Neat little circuit. Lifting the source of the FET's from ground (adding diodes there,) is "somewhat like" the same idea as a cathode resistor used to create bias for a tube. Two things that make FET's hot: Slow switching times (for a single FET,) and no "dead time" between the transition from one FET to the other. Vibrators have a long dead time in the transition as just a natural effect of the moving reed. Picture this: Think of the moving or swinging reed in the middle of the vibrator contacting one side, then moving to the other contact on the other side "as it's vibrating." The reed (as its moving from one side to the other) for a brief period "when its between contacts" has no contact to either side (this is "dead time.") so effectively, the primary of the vibrator transformer is completely open for a small amount of time. This dead time allows the magnetic field of the transformer to collapse successfully before switching on the other FET at the other end of the winding. The CD4047 is a great choice for a low voltage oscillator (6V car) but has an immediate transition from one FET to the other with no dead time, that alone will make the FET's heat. This circuit could benefit by a delay line in the gate path of the FET's, or some other way of creating dead time. Dead time is easy to make, but many circuits will slow the IC's output rise time drive signal as the FET is turned on, this will also create heat as the FET is now operating for a longer period in the linear region. The TL494 has a "dead time control" pin, but the downfall is: the IC will only operate down to 7 Volts, then it looses control over the feedback ETC... (Great for 12V car Vibrator circuits though!) Building a small auxiliary switching supply for a TL494, "so it can operate safely with a supply down to about 3V" would be the optimum vibrator circuit for a 6V car. This could be done with a simple boost converter using an inductor. Brute force over rated "lower logic level" gate drive FET's with a low on resistance will make a circuit like the one shown here work fine, but the transformer will get warmer due to the lack of dead time. "You move the problem on to the next stage, so to say :^)" If you made it this far reading this, you are probably "really into electronics :^)" Have a good day, and thanks for taking the time to make your videos Shango, we all enjoy them a lot! Note: Edited due to spell check :^/
Hello Paul and yep I read it all, comprehend? maybe. Thanks for commenting here. So if we think about applying the TL494 dead time control pin, we would boost the 3V to the 7V the IC needs? Saw elsewhere here concern with noise using a DC/DC converter, to come up from the 3V. Do you see that as significant compared to the other switching going on on the board? Thanks
@@PracticallyFixed Hi. No, dead time just puts a delay in the transition from one FET to the other. Another small boost supply involving an inductor would need to be used to lift the 6V supply up to 12V, only for the TL494, (and hold it there). The boost supply would need to operate with an input voltage down to about 3V (worst case.) A 6V car sitting off without the engine running (no generator or alternator spinning) can have a voltage at the radio as low as 4V, the boost circuit should have a Volt "or so" safety region.
Awesome. I really want to try the 3525 with the duty cycle control as an experiment. My experience is any of those buck boost converters generate so much noise they just drive am radios completely nuts so I'm trying to avoid that type of voltage manipulation
@@MrCarlsonsLab Right, I think I got that. So in principle we would leave the FETs as they are, use the TL494 providing PWM instead of using the CD4047 (or 3525) for driving the FETs because it can provide the dead time control needed for a happy transformer? If the supply is, say as low as 4V, use a DC/DC boost to just supply the Vcc (12V) for the TL494 driving the FETs (which would still be switching the 4V from the transformer center tap to ground via the different legs). If so, then Shango would have better luck with the TL494 than the 3525? I checked a datasheet for the 3525 and did not see a dead time control capability but maybe I missed it (correction, yes pin 7 allows dead time adjustment). Thanks both of you. Edited due to misspelling the CD4047 and correction to dead time capability to 3525.
@@shango066 I used a DC/DC converter off Amazon in buck mode once in an experiment to replace a bias cell in a 1937 Grunow radio once (was actually testing to see if the bias cell voltage, 1V vs 3V, was critical) and didn't pick up any noise - maybe I was just lucky? Also thinking about the steel shielding you have around the vibrator socket? I had someone knowledgeable in RF question me once about noise from a SS vibrator in a radio I did last fall about RF noise - all I could say was that I didn't get any, and it also could have been due to the steel partitions in the chassis. Maybe the reactance of the transformer chokes if off the power line as well? Thanks, great video.
Years ago, I had a 6 volt CB radio in my 12 volt car. I bought a dropping resistor, but it did a poor job, as the radio drew a large current when transmitting, and burned out the resistor. Rather than using a divider resistor, I would build a 6 volt voltage regulator. That way, if the current draw changes, then the regulator would supply it without fear of burning out. At the time I had that radio, it had a solid state vibrator installed, and it worked very well and very quietly. As a kid, I really loved the old car radios with tubes! While harder to work on because of the compactness of the units, they were well built because of the purpose of the radio. The failed 0.5 uF caps were not hurt by the DC voltage, but by the voltage spikes from the vibrator switching, so the voltage rating being higher would be better than if you drop the voltage rating!
@@Rev22-21 Are you speaking of the dropping resistor? If so, the resistor is a low ohm unit with a tap in the center of it, and it draws many amps that are wasted in heat. When you add your tube radio to it, it makes the resistor draw even more power, and it creates more heat! Mine got so hot, you could not touch it, and it burned out because of all the heat.
In 1977 I was given a 1952 Buick Special straight eight automatic. My Grand Mama’s car since new. It was 6 volt and we converted it to 12 volt. The radio was beautiful and sounded amazing. So I found a similar 12 to 6 volt converter on J C Whitney for like $10.00. Similar design in ceramics. Mounted it on a steel under dash point and it ran warm to touch. The heater motor was not changed and ran fast but never failed.
This has been a great series. Thanks for being so quick in getting it on UA-cam. Try a 10 watt 250 ohm resistor in series with the B+. It will simulate a hollow state rectifier like a 5U4. They pretty much have the same characteristics. I've had my share of pesticides today! And lead solder too! No vendetta COtrol though!
I think the schematic was saying you could put diodes in the negative rail (the combined one that both the fet sources and chip are powered from), that would mean the chip ground is still firmly tied to the fets source and you wouldn't get problems like you did. But you would have a problem with the gate voltage being too low to turn on the fet strong enough (4V may even be too low to turn on some fets at all). It's better if you keep the ground solid and put diodes just on the transformer side like you did. A problem with this circuit is it has zero deadtime. There is some deadtime added by the resistors and diodes to the fet gates (turn-on slower than turn-off). A SG3525 or TL494 (KA7500) is a much better choice for this application. Though as you said they might not work at 6V, you'd need to check the datasheets. An idea I have is to use a 4017, connect output 1 to gate 1, 3 to gate 2, leave outputs 2 and 4 disconnected, connect 5 to reset. This will give you a 50% duty cycle. But you need another chip to generate the clock for the 4017.
Another idea, you could put diodes between drains of fets and transformer. Thus you wouldn't need to modify the radio. symmetrical (same amount of diodes on each transformer leg), just means you need twice the number of diodes than if you put the diodes on the transformer center tap or on the negative rail of the circuit. But the advantage of not putting diodes in the negative rail is you can feed a full 6V power to the chip which will give you the full voltage on the fet gates which will lower their on resistance and make them run cooler. You would only need to run one extra wire from 6V power to the chip power. Regarding the 12 to 6V dropping resistor. What's the current draw of the radio? Maybe you could use a linear reg like a LM317 with added transistor to increase its current. The datasheet has good examples, like added transistor plus current limiter to protect it.
Yes, diodes in the source lead is bad. The potential Vgs (from the driver) needs to be > the miller voltage not just the max threshold (typ 4v). There is quite a lot of theory and practice going into driving a fet. Those diodes added in the source lead to lower the output would mean those fets are not being switched efficiently thus leading to losses in the fets so they get hot. Edit: pulled up the datasheet for the IRF3205. Vgs_th is 2-4V and the miller voltage is about 5.2v from the charge plot. So, those fets will not be fully enhanced unless Vgs is at least that much.
A thought on the sideline: If the circuit must be for a 6v radio, it may be a plan to boost the chip supply voltage via a low frequency boost converter to ensure the FETs are fully conductive during their duty cycle.
Blitz Roehre you could derive it from two diodes, one from each leg of the transformer. Once the converter starts up, the transformer leg that's not pulled to ground will be at 2*VBatt. So it will derive its own doubled supply voltage. The circuit must still somewhat work just to turn on the fet enough that it pulls one leg of the transformer down. Vgs threshold must be less than 6V. After one cycle it will already have 12V supply.
Great that you got the original set working. I have 3 delco sets in my old gm cars. Pity no AM stations around any more.Too many sets to modify for rediwhatever.
Excellent video; love the experimentation. "you built it yourself" - amen. Yes, I have used off the shelf SS vibrators but would love to make one myself and perhaps modify for different applications I have had that require a different voltage, e.g. 4V, plus "I built it myself". Also really enjoy the comments and suggestions by the other knowledgeable folks here for other ideas. If there are other schematics to look at (other than those Shango showed) I would like to see them if either Shango can leave a link or those who sent them in to him would be so kind. Thanks
Hi, adjusting the switching frequency to a value closer to the original vibrator's may benefit the audio by reducing the hum. Using SG3525 instead of CD4047 allows you control the output voltage more or less at will, since you can adjust the duty cycle. By adjusting the duty cycle you can adjust the magnetic flux of the transformer (the same you can do by adding resistors or diodes to the primary with the fix 50% duty cycle CD4047). An additional benefit of the 3525 is that you can add dead time to the gates driving to make sure that the transistors don't overlap in conduction reducing heat ( but at so low freq I don't thing that is a big deal). Best wishes from Buenos Aires, Argentina.
I had a generic tractor/truck radio originally sold at princess auto that ran six or 12 volt positive or negative ground. To run 12 volt you attached a big ceramic resistor in the supply line. It hung off the radio and got very hot. It worked great, wish I'd kept it.
Yep,all my Chrysler products have them. You always carry a spare with you,just in case. And like Shango pointed out,you can burn yourself touching them.
@@Suddenlyits1960 Mount a tincan designed to hold a coffee cup to the transmission tunnel. With the bleeder resistor in the bottom to act as a heating coil to keep your coffee warm. Sorry I have an odd sense of humor.
One radio I did 6 to 12 I separated the filaments into two 6v groups then series them together. I had to go back and rebalance/rewire to get the current right one tube ran hot. I used the cheap ebay voltage gen. 12 to 220 but got lucky with a quiet one, might have ran it through the transformer as a choke.
On some of those schematics found on the internet the one FET turns off just as the opposite one becomes conductive. This basically means the magnetic field in the transformer core is being force-reversed, a lot more current flow occurs through the FETs here generating more heat. Better is a 80% duty cycle setup which has pause between each switch on to allow the magnetic field in the trafo-core to decay. A microcontroller programmed to do this would be the ideal solution. Also, the use of IGBTs designed specifically as automotive ignition coil drivers would make the circuit very resistant to voltage spikes etc. Fairchild has coil drivers in their product portfolio. I am working on a setup like this and will upoad a video if it works out ok..
I enjoy watching you work on these. Consider me a subscriber seems radio guys are a dying breed so always looking for more people to learn from. Also am loving the hot takes on the different audio that comes through.
What appears to have been overlooked is that there was a capacitor in parallel with one winding of the transformer, forming a tuned circuit with the transformer inductance. When both vibrator contacts to the primary opened, the tuned circuit would start a decaying oscillation. After a half cycle, the polarity on the transformer windings has reversed. The vibrator was designed to connect its other contact as close to the voltage peak as practical within tolerances. There is a short time when both contacts are open, just long enough for the capacitor to reverse polarity. The voltage difference across the contacts, just before they closed, was minimal, greatly reducing arcing and extending the life of the contacts. The energy stored in the transformer inductance is not lost. It is temporarily stored in the capacitor and returned to the transformer, but with the transformer's magnetic field reversed. If the capacitor is removed and the FETs do not provide an interval when both are open, the energy stored in the transformer's inductor will be dissipated in the FETs and converted to heat. In the vibrator design, If the capacitor were bad or off value, the vibrator contacts would have a short lifetime, due to arcing. Replacing the vibrator would get the radio working again, but the contacts on the new vibrator would burn out quickly and the new vibrator would fail. The vibrator was socketed for simple swap out. If the radio was burning out vibrators, it needed to be removed and checked for failures of the capacitor or other components in the power supply. Vibrator power supplies in car radios were used until about the mid 1950s, when power transistors for the radio's audio output stage were developed. Then, "space charge" vacuum tubes, which operated from a 12 volt B+ supply, were used for all the radio's circuitry except the audio output stage, which had the only tube which required the higher B+. The expensive, heavy, vibrator power supply was designed out, and the art of designing them was lost. The "space charge" radios were in production for only a few years before being superseded with all transistor radios. However, car radios were produced in high volume, so millions of "space charge" radios were produced before the design was removed from production.
I'm fairly certain that those vibrators ran at 115 HZ. You don't want ot run much lower as the transformer will draw more current due to core saturation with the lower frequency.
JAMES AZBELL I am from Ohio and I have brother in Birmingham AL. Iwould use a microprocessor solution to that vibrator replacement, but you have done a great job with analog.
I actually did something like this a few years ago on a Motorola radio. here is a list of the things I did. Instead of using the CD4047. I used a TL494 chip. This is a plain vanilla generic PWM driver chip that can be found in about 90% of the computer power supplies and high powered car amps out there. The nice thing about using this chip is I can now have voltage regulation. The B+ will stay constant regardless of the voltage on the car's electrical system. To run the vibrator step-up transformer at 12 volts. I increased the frequency from 115Hz to 340Hz. The reason for doing this is to prevent the core of the transformer from saturating. I also replaced the rectifier tube with solid state diodes. As far as the tube filaments go. I used both a dropping resister and a linear voltage regulator. I used the dropping resister to bring the voltage from 12 volts down to 8 volts. Then I ran the 8 volts into a 7805 linear voltage regulator with the ground pin floated with 2 1n4001 diodes. this gave a regulated output voltage of 6.2 volts. The reason for pre-dropping the voltage with the dropping resister was so the VR didn't have to brunt the full voltage. the VR only had to handle a 2-3 volt drop instead 6-8 volts and therefore, runs cooler.
Just put a lm317 voltage regulator on your board and adjust it to proper voltage out from the transformer. That lets you ditch the diodes and the zener’s.
Very Sad to hear about the "This is a test" girl". I will miss her. I hope the Zombies didn't get her ? Excellent Video , We use to take the old vibrators apart and gently sand the contact points. We would then put a tiny drop of Transmission or power steering fluid on those points. I had them last as long as 5 more years. ( our old radio was used daily ) ( 1954 Ford radio) The old 0Z4 tubes were in just about EVERY GM radio back then. The old Fords ( Bendix or Zenith ) used 6X5 rectifiers. The tubes were cheap & plentiful. Few folks fix these old radios these days. They just don't even bother teaching this old technology in school today. Thank You for yet ANOTHER Excellent Video.
Product design versus electrical needs is always the toughest nut to crack. The offerings by sellers of solid state vibrator replacements offer no guarantees. Makes sense after reading Paul Carlson's comment.
12:15 "This is why I don't like working on stuff in the house" Agreed. Especially when you notice specks of Dust moving. Because you have just introduced House Dust Mites into your home.
What I find impressive (besides how well this ended up working) is that all that was needed to replace a mechanical vibrator. Score one for old school!
The way I would do it is rewiring the transformer, that is, just use two IR2153 as a full bridge driver and connect the whole primary across it. Heater wise, make a single transistor running at low duty cycle, switching the heaters on and off from 12V.
No need to add that masive resistor to drop B+, you can add a series variable resistor to that 22k on pin 2 to adjust the frequency. Also if both fets get hot it could be because the output of the 4047 might be turning both fet at the same time similar to the crowbar state on a H-bridge of a switchmode power supply. You can try adding a delay on one of the fets or use an oscilloscope to see how they are switching. The gate of the FET is a capacitor which means that if the output of the 4047 is not pulling it to ground fast enough it will create this problem. Or you could change one of the FETs gate resistor to a lower value. That way you'll be adding a delay.
The minimum frequency of that chip is 100 Hertz so what am I going to do speed it up to the point where it's out of the Transformers response? I can see adjusting the duty cycle yes not the frequency
@@shango066 Since the goal is to reduce the output voltage, tuning it to a less efficient point in the transformer's response curve may be a reasonable way to accomplish that as long as the transformer core doesn't get too warm. Alternately, 555 is the classic way to do a variable duty cycle low freq square wave, and it's easy to make an inverter to get the other phase gate drive using a transistor and a resistor.
@@shango066 I dont have a 4047 to test, but I did a simulation using proteus and changed the timing cap for a 220nF and a 22k resistor and I have a frequency of 42Hz
Monostable multivibrator circuits are also useful in creating a nice sounding click from the weak pulse output of some vintage geiger counters like Eberline and the Civil Defense CDV700. You can find some schematics online if interested. Their main benefit is super low battery consumption so one battery can last for years if usage is occasional and you remember to turn the power off.
You and Bob Anderson should do some sort of vintage electronic event. Add Mr Carlson, radioTVphonorant, Jordan Pier, MrRadio, Buzz1151 and Glasslinger.
The diodes are dropping the voltage to the oscillator chip and hence dropping the gate drive to the fets. If you run the oscillator chip on the full 6 volts to maintain full gate drive, all should be well. Good luck with the project and thank you for your entertaining and education videos over the years. On reflecion, it would make more sense for the diodes or resistors to connect to the centre tap of the primary and leave the drive circuitry at full power.
That's one reason why I like the idea that Jernej Jakob / braindamage had, of putting the dropper diodes in the two legs of the transformer primary, rather than on the center tap. That keeps up the voltage to the chip, and the FET gates. One downside to that method, though, is that when the primary rings, instead of the back-voltage going harmlessly through the built-in snubber diodes in the two FETs, the series dropper diodes will block the backflow, which could lead to those series diodes seeing some hefty reverse spikes (depending on how fast the FETs are switching on). Simple series dropping resistors in the two primary legs (instead of diodes) could avoid those problems. The thing is, since Dan doesn't want the RF noise of a boost-buck converter, you're going to be dropping voltage somewhere, with diodes or resistors, and that means heat, so he has to choose where to locate those components so that they don't cook the little FET board.
@@auchterawer1150 Further thoughts: I can't see that the FET drive will produce any more B+ than the origial vibrator with it's metal contacts. 350V may be a bit on the high side but if it doesn't exceed the tube or capacitor ratings why worry?
Well done Sir! I guess you can't adjust the Mark to Space or Duty Cycle as with a 555? It seems as though this circuit would really benefit from having a Set H.T. Preset.
I think the heating encountered when putting the diodes between the supply and the center tap is caused by the inductance of the transformer. When one side turns off the current wants to continue flowing so the voltage shoots up. Normally the current continues by flowing backwards through the other winding, but with the diode present it can't. So the voltage shoots up as the transistor turns off and causes excess heat.
Putting the diodes where they suggested would lift the source voltage above ground effectively lower the difference between gate and source, the fets would probably have been running in their linear region
I really appreciated and enjoyed this video on building the solid state vibrator circuit, great job there Shango066. I have a small collection of vintage tube type auto radios. So far, I've been successful opening the vibrator "can" and cleaning up the contact points, and reassembling them to work. Someday I will build one of these SS vibrators and try it out! On another subject...I wonder what happened to your "this is a test, this is only a test" radio station? I think you mentioned before that it had been on a long time. I wonder who was broadcasting that signal and who was paying for it.
It does not like 2 diodes because you are lowering the voltage of the CD40XX chip also and it needs at least 5 volt or more to work correctly and with an input of 6 volts and 2 diodes in series with the circuit the CD40XX chip is not getting the required 5 volts to work. You should change the R/C components around the chip to obtain the desired output voltage instead of adding diods to drop the voltage. Or you could use the NE555 8 pin chip to do the PWM for the mosfets and it would be totally able to regulate the voltage the way you want by using a little pot to do that.
Fet's have a gate threashold voltage, commonly 4v, if the power supply drops to around this voltage the fets will go into linear mode due to low drive voltage, and will get hot.. Bipolar trannies may have been better at 6v power. I'd like to try this with a royer oscillator, the o/p is close to a sine wave, and the circuit is simpler.
That .004 capacitor on the secondary is probably to reduce ringing of the transformer, you see caps like that across the secondary in many tube radios with solid state rectifiers to reduce ringing caused by the slow and abrupt recovery of solid state diodes. that cold cathode rectifier tube may have had a similar probably so they put that capacitor there.
Sg3525 would have been a good option you would need a boost converter/charge pump or logic level fets but ive got a feeling the sg3525 will want 12-15v it has UVLO and soft start to protect the fets it's a great ic fully adjustable frequency duty cycle dead time etc also a powerful fet output so you can drive fets directly, it virtually a tl494 but with a nice fet output stage , it has 2 error amps so easy regulation.. its not pin compatible with the tl494 however.
With a radio that needs 220V, anyone who doesn't want to do much soldering and doesn't mind having something conspicuously 21st Century under the hood could just order a 12-to-230V power inverter intended for a high voltage country like the UK or Australia. I see a bunch of them on Google from $20 to $500, probably depending on how much Chinesium you can tolerate.
if you used an H bridge across both windings of the transformer at once (instead of the push-pull) it should be possible to run it on 12V since you're using both windings at once instead. but of course that means an H bridge driver instead of just two fets. I don't think the transformer would like 12V across each winding even when PWMing it since it would cause it to saturate.
Wouldn't you have to separate the two primaries to do that? And then too...wouldn't firing them simultaneously 'buck' each other and cancel each other out?
@@Rev22-21 the primary is center tapped, so the two coils are actually out of phase with each other. this is how it works as a push-pull. If both windings were in phase it wouldn't push-pull any more. By not using the center tap and leaving it free floating, this converts it effectively into a single winding instead. You need to drive it with an H bridge though, because doing this loses the benefit of having the center tap to reference from. It's similar to how you can make a full wave rectifier with 2 diodes and a center tapped transformer, or use a single winding and a 4 diode bridge.
By the way, it appears that those black capacitors at 14:00 *_are_* already marked as to the inside/outside of the foil with the little nub by one pin...
To drop the 12v to 6v for the radio, maybe a 6.2v high wattage zener diode in series with the power lead to the radio would work better? Actually, using the ''3525 chip and replacing the tubes with their 12 volt versions (14xxx series) might be the best idea. The transformer is a 6v one, but you could probably run it on 12v by cranking down the duty cycle and putting diodes in series with the power input to the inverter. Then you wouldn't need that big ass resistor! (Yeah also have to replace the pilot lamps too). Back in the '70s, a friend of mine built a DC-DC inverter for a ham mobile rig using a power transformer with a 12vct filament winding being feed by a pair of cross coupled 2N3055 transistors. The filament windings were used as a push pull primary, and feedback was crude, power resistors from the base of one transistor to the collector of the other, capacitors in parallel with the resistors. We played around with the values to get the best efficiency and output voltage. The HV secondary was used as normal, the primary was used to generate bias voltage for final amp tubes in the rig. The rectifier filament winding wasn't used for anything (NC). The thing worked, but it was sorta Rube Goldberg.
When I was installing a 12 V negative ground stereo in my 12 V positive ground Triumph, I had an inverter-like device I got years ago. It could also be wired to produce 12 v from 6 volts and vice versa depending on how it was wired. But I haven't seen anything like it lately. It came from Radio Shack. Probably would have been simpler just to convert the car to negative ground.
@@eeengineer8851 Ah, never considered vibration. I have a Granco FM-only table radio that's slug tuned which performs very well. I never figured out why they went with that instead of a variable capacitor either.
@@defaultuserid1559 tuning capacitor for only 88 - 108 MHz is expensive for what it is. TVs were using slug tuning and then went to a cheaper form capacitor tuning for a while before they went to varicap diode.
I have seen rectifier tube deletes using modern Silicon rectifiers using series resistors to lower the B+ Voltage. This would lower the regulation but no more than the original transconductace value of the tube. The current drain should be "constant" providing the audio output is class A. The AGC may affect it slightly though.
Normal running voltage should be about 7 to 7.2 volts. Of course old generators put out nothing at idle so you could go down to 6 or even a little lower with lights and heater on. Car radio tubes were supposedly designed to operate with the varying voltages of old fashioned poorly regulated generator systems, although I think any indirectly heated tube is pretty tolerant of varying heater voltage.
Well done on the vibrator mod. You could try replacing the 22k timing resistor with a pot (say 100k) and use that to change the oscillator instead of using diodes,' but, needless to say, test out of circuit (with test transformer) first.
I need to replace my vibrator. Is it worth getting the replacement or is the solid state a better option? Would you make me one if I paid you? I’ve replaced all the tubes with no luck.
I wonder if you could get away with a ZVS and ferrite transformer, the output on the secondary is sinusoidal, run the gate supply directly from the 12-14v and use a BJT pass element on the input to the fets to regulate the output B+ on the secondary, I think you might be able to get away with it with a bit of filtering etc, it would preserve the original transformer too
needs short circuit protection. Current sense resistor and a transistor to disable the MV. No need for the 1 00000uf crapacitor there. remember the Royer circuit uses a impedance, an inductor in the B+. The Diodes in the source reduces the drive too much already with 6volt
Risking a repeat here (of what I asked of Mr. Carlson) has anyone considered using a PIC in place of the IC? An 8 pin PIC could be easily programed to control those FET's. A few lines, a little time and presto....
@@shango066 : The following YT Channel does a pretty good job at explaining the A-Z's on PIC's . The Channel is simply called Learning Microcontrollers. I've built two major projects myself; one (a one year long project) was patented and one was recognized & published in Coca-Cola USA Tech magazine. The former started out as a logic IC topology, went through seven board upgrades and finally into the PIC Microcontroller arena. I never looked back.
@@shango066 Learning how to program is like learning how to play a musical instrument, rebuild a carburetor or put a puzzle together. You watch others "program", repeat what they do and before you know it you're making an LED blink, making a switch turn on/off something, building a timer of sorts, monitoring sensors and then building thermonuclear war head (not really). But you get the idea. 😜 It's like moving from discrete components to IC's. Simple.
Maybe try two rectifier diodes in series then add a cap to ground at the centre tap to average out the ripple voltage, it might bring it in to the sweet spot and also take some stress off the diodes?
By putting diodes in the source you raise the source to 1.2V, and then probably the MOSFET is not opening full = heat. I think you had that figured out the first build. You measured the gate as it was a bit above 2V, and that will stay. Just put a few diodes in the positive, as (i think) you did the first time ? You can put as many as you need.
cant you change the freqenty of the vibrator board to lower te output voltage? in that case yo dont need the diodes.. (its a pwm chip, so mounting a potentiometer on it would make it adjustable for the input voltage of 7 volts.whitout extra heat...)
What about rebuilding that set using the tubes designed in 1957 that has a H+ of 12V and a B+ of 12V and using the two VMOS FETS as Transformerless Class A output stage to drive the speaker.
@shango066 Trippin' on a hole in a paper heart. If the trip was caused by a wet floor contaminated with Camp Lejeune water you might be entitled to compensation.
I read about you are not able to se a DC-DC .... sure, those are noise sources. You can use an LM350 mounted directly on the case, (use the whole thing as a heatsink). Not the pretiest .... Not sure if the 350 will like the load at all. It is a shame because you need to modify the radio a bit. At least you can get some control. Be sure not to starve the IC or be to much low that Vgs on the mosfets are not high enough so they wont turn into heat sources.
Imagine trying to replace a Dynamotor power supply for the OLD MOTORONI police radios! NOT! Can you imagine what the factory charged for the "radio option" in the day?
@@shango066 Hmm.. they can be noisy, but sticking it in a metal can and filtering the input and output would probably sort that out. Might be worth a try if you have issues with that space heater of a resistor - would give better DC regulation.😀
3:05 I'm disappointed! I was hoping to see you modify the radio with 12-V-plate space charge tubes then build a one-transistor output stage with a giant heat sink. It's important to keep up with the times, you know.
Ye Gods ! Electronics from the Middle Ages! Talk about crude and primitive. To get the final output voltage regulated don't have series diodes and dropper resistors as you'll be faffing with them forever. Add a voltage control loop instead. Pot-down a sample of the B+ output voltage and when it goes over a set threshold turn off the FET drive. Can use a Vbe junction as a comparator. The circuit then runs in squeg mode. Does the B+ voltage really need regulation anyway? Valves (tubes) and good capacitors can tolerate huge variations in anode voltage. The 12V supply input ranges from 11.5V to 14.8V anyway. EDIT: the second schematic with the generic SMPS IC , it's being used here 'open loop', normally is used closed loop as I describe above. This chip has a built in switching crossover function to prevent both FET conduct at same time, I think it's 2% though it's been 30 yrs since last used this chip.
why not to replace the vibrator and the rectifier tube and use an external 12 to 6 v converter? As a radio collector I do not like this solution , I like the video anyway
Neat little circuit. Lifting the source of the FET's from ground (adding diodes there,) is "somewhat like" the same idea as a cathode resistor used to create bias for a tube. Two things that make FET's hot: Slow switching times (for a single FET,) and no "dead time" between the transition from one FET to the other. Vibrators have a long dead time in the transition as just a natural effect of the moving reed. Picture this: Think of the moving or swinging reed in the middle of the vibrator contacting one side, then moving to the other contact on the other side "as it's vibrating." The reed (as its moving from one side to the other) for a brief period "when its between contacts" has no contact to either side (this is "dead time.") so effectively, the primary of the vibrator transformer is completely open for a small amount of time. This dead time allows the magnetic field of the transformer to collapse successfully before switching on the other FET at the other end of the winding. The CD4047 is a great choice for a low voltage oscillator (6V car) but has an immediate transition from one FET to the other with no dead time, that alone will make the FET's heat. This circuit could benefit by a delay line in the gate path of the FET's, or some other way of creating dead time. Dead time is easy to make, but many circuits will slow the IC's output rise time drive signal as the FET is turned on, this will also create heat as the FET is now operating for a longer period in the linear region. The TL494 has a "dead time control" pin, but the downfall is: the IC will only operate down to 7 Volts, then it looses control over the feedback ETC... (Great for 12V car Vibrator circuits though!) Building a small auxiliary switching supply for a TL494, "so it can operate safely with a supply down to about 3V" would be the optimum vibrator circuit for a 6V car. This could be done with a simple boost converter using an inductor. Brute force over rated "lower logic level" gate drive FET's with a low on resistance will make a circuit like the one shown here work fine, but the transformer will get warmer due to the lack of dead time. "You move the problem on to the next stage, so to say :^)" If you made it this far reading this, you are probably "really into electronics :^)" Have a good day, and thanks for taking the time to make your videos Shango, we all enjoy them a lot! Note: Edited due to spell check :^/
Hello Paul and yep I read it all, comprehend? maybe. Thanks for commenting here. So if we think about applying the TL494 dead time control pin, we would boost the 3V to the 7V the IC needs? Saw elsewhere here concern with noise using a DC/DC converter, to come up from the 3V. Do you see that as significant compared to the other switching going on on the board? Thanks
@@PracticallyFixed Hi. No, dead time just puts a delay in the transition from one FET to the other. Another small boost supply involving an inductor would need to be used to lift the 6V supply up to 12V, only for the TL494, (and hold it there). The boost supply would need to operate with an input voltage down to about 3V (worst case.) A 6V car sitting off without the engine running (no generator or alternator spinning) can have a voltage at the radio as low as 4V, the boost circuit should have a Volt "or so" safety region.
Awesome. I really want to try the 3525 with the duty cycle control as an experiment. My experience is any of those buck boost converters generate so much noise they just drive am radios completely nuts so I'm trying to avoid that type of voltage manipulation
@@MrCarlsonsLab Right, I think I got that. So in principle we would leave the FETs as they are, use the TL494 providing PWM instead of using the CD4047 (or 3525) for driving the FETs because it can provide the dead time control needed for a happy transformer? If the supply is, say as low as 4V, use a DC/DC boost to just supply the Vcc (12V) for the TL494 driving the FETs (which would still be switching the 4V from the transformer center tap to ground via the different legs). If so, then Shango would have better luck with the TL494 than the 3525? I checked a datasheet for the 3525 and did not see a dead time control capability but maybe I missed it (correction, yes pin 7 allows dead time adjustment). Thanks both of you. Edited due to misspelling the CD4047 and correction to dead time capability to 3525.
@@shango066 I used a DC/DC converter off Amazon in buck mode once in an experiment to replace a bias cell in a 1937 Grunow radio once (was actually testing to see if the bias cell voltage, 1V vs 3V, was critical) and didn't pick up any noise - maybe I was just lucky? Also thinking about the steel shielding you have around the vibrator socket? I had someone knowledgeable in RF question me once about noise from a SS vibrator in a radio I did last fall about RF noise - all I could say was that I didn't get any, and it also could have been due to the steel partitions in the chassis. Maybe the reactance of the transformer chokes if off the power line as well? Thanks, great video.
Years ago, I had a 6 volt CB radio in my 12 volt car. I bought a dropping resistor, but it did a poor job, as the radio drew a large current when transmitting, and burned out the resistor.
Rather than using a divider resistor, I would build a 6 volt voltage regulator. That way, if the current draw changes, then the regulator would supply it without fear of burning out.
At the time I had that radio, it had a solid state vibrator installed, and it worked very well and very quietly.
As a kid, I really loved the old car radios with tubes! While harder to work on because of the compactness of the units, they were well built because of the purpose of the radio.
The failed 0.5 uF caps were not hurt by the DC voltage, but by the voltage spikes from the vibrator switching, so the voltage rating being higher would be better than if you drop the voltage rating!
Good idea. If it were on the dash and away from the radio....why not?
@@Rev22-21 Are you speaking of the dropping resistor?
If so, the resistor is a low ohm unit with a tap in the center of it, and it draws many amps that are wasted in heat. When you add your tube radio to it, it makes the resistor draw even more power, and it creates more heat!
Mine got so hot, you could not touch it, and it burned out because of all the heat.
In 1977 I was given a 1952 Buick Special straight eight automatic. My Grand Mama’s car since new. It was 6 volt and we converted it to 12 volt. The radio was beautiful and sounded amazing. So I found a similar 12 to 6 volt converter on J C Whitney for like $10.00. Similar design in ceramics. Mounted it on a steel under dash point and it ran warm to touch. The heater motor was not changed and ran fast but never failed.
This has been a great series. Thanks for being so quick in getting it on UA-cam. Try a 10 watt 250 ohm resistor in series with the B+. It will simulate a hollow state rectifier like a 5U4. They pretty much have the same characteristics. I've had my share of pesticides today! And lead solder too! No vendetta COtrol though!
It’s so cool when the factory radio works in a vintage vehicle. The radio in my old ‘58 Chevy still works, and I use it almost everyday.
I think the schematic was saying you could put diodes in the negative rail (the combined one that both the fet sources and chip are powered from), that would mean the chip ground is still firmly tied to the fets source and you wouldn't get problems like you did. But you would have a problem with the gate voltage being too low to turn on the fet strong enough (4V may even be too low to turn on some fets at all). It's better if you keep the ground solid and put diodes just on the transformer side like you did. A problem with this circuit is it has zero deadtime. There is some deadtime added by the resistors and diodes to the fet gates (turn-on slower than turn-off). A SG3525 or TL494 (KA7500) is a much better choice for this application. Though as you said they might not work at 6V, you'd need to check the datasheets. An idea I have is to use a 4017, connect output 1 to gate 1, 3 to gate 2, leave outputs 2 and 4 disconnected, connect 5 to reset. This will give you a 50% duty cycle. But you need another chip to generate the clock for the 4017.
Another idea, you could put diodes between drains of fets and transformer. Thus you wouldn't need to modify the radio. symmetrical (same amount of diodes on each transformer leg), just means you need twice the number of diodes than if you put the diodes on the transformer center tap or on the negative rail of the circuit. But the advantage of not putting diodes in the negative rail is you can feed a full 6V power to the chip which will give you the full voltage on the fet gates which will lower their on resistance and make them run cooler. You would only need to run one extra wire from 6V power to the chip power.
Regarding the 12 to 6V dropping resistor. What's the current draw of the radio? Maybe you could use a linear reg like a LM317 with added transistor to increase its current. The datasheet has good examples, like added transistor plus current limiter to protect it.
If you close the gate off the cows and sheep won’t get out. Works for me
Yes, diodes in the source lead is bad. The potential Vgs (from the driver) needs to be > the miller voltage not just the max threshold (typ 4v). There is quite a lot of theory and practice going into driving a fet. Those diodes added in the source lead to lower the output would mean those fets are not being switched efficiently thus leading to losses in the fets so they get hot.
Edit: pulled up the datasheet for the IRF3205. Vgs_th is 2-4V and the miller voltage is about 5.2v from the charge plot. So, those fets will not be fully enhanced unless Vgs is at least that much.
A thought on the sideline: If the circuit must be for a 6v radio, it may be a plan to boost the chip supply voltage via a low frequency boost converter to ensure the FETs are fully conductive during their duty cycle.
Blitz Roehre you could derive it from two diodes, one from each leg of the transformer. Once the converter starts up, the transformer leg that's not pulled to ground will be at 2*VBatt. So it will derive its own doubled supply voltage. The circuit must still somewhat work just to turn on the fet enough that it pulls one leg of the transformer down. Vgs threshold must be less than 6V. After one cycle it will already have 12V supply.
Great that you got the original set working. I have 3 delco sets in my old gm cars. Pity no AM stations around any more.Too many sets to modify for rediwhatever.
Excellent video; love the experimentation. "you built it yourself" - amen. Yes, I have used off the shelf SS vibrators but would love to make one myself and perhaps modify for different applications I have had that require a different voltage, e.g. 4V, plus "I built it myself". Also really enjoy the comments and suggestions by the other knowledgeable folks here for other ideas. If there are other schematics to look at (other than those Shango showed) I would like to see them if either Shango can leave a link or those who sent them in to him would be so kind. Thanks
Hi, adjusting the switching frequency to a value closer to the original vibrator's may benefit the audio by reducing the hum.
Using SG3525 instead of CD4047 allows you control the output voltage more or less at will, since you can adjust the duty cycle.
By adjusting the duty cycle you can adjust the magnetic flux of the transformer (the same you can do by adding resistors or diodes to the primary with the fix 50% duty cycle CD4047).
An additional benefit of the 3525 is that you can add dead time to the gates driving to make sure that the transistors don't overlap in conduction reducing heat ( but at so low freq I don't thing that is a big deal).
Best wishes from Buenos Aires, Argentina.
I like the 12V to 6V dropper resistor idea. It makes things so much easier. :)
I had a generic tractor/truck radio originally sold at princess auto that ran six or 12 volt positive or negative ground. To run 12 volt you attached a big ceramic resistor in the supply line. It hung off the radio and got very hot. It worked great, wish I'd kept it.
I like that it's a period correct part too. I'd mount that resistor on the firewall under the hood so it's visible. Super cool.
Chrysler products used to take a ballast resistor and drop the 12v to 6v for the coil. You can still buy them.
Yep,all my Chrysler products have them. You always carry a spare with you,just in case. And like Shango pointed out,you can burn yourself touching them.
@@Suddenlyits1960 Mount a tincan designed to hold a coffee cup to the transmission tunnel. With the bleeder resistor in the bottom to act as a heating coil to keep your coffee warm. Sorry I have an odd sense of humor.
One radio I did 6 to 12 I separated the filaments into two 6v groups then series them together. I had to go back and rebalance/rewire to get the current right one tube ran hot. I used the cheap ebay voltage gen. 12 to 220 but got lucky with a quiet one, might have ran it through the transformer as a choke.
Thank you for making a video of this; very informative and entertaining
On some of those schematics found on the internet the one FET turns off just as the opposite one becomes conductive. This basically means the magnetic field in the transformer core is being force-reversed, a lot more current flow occurs through the FETs here generating more heat. Better is a 80% duty cycle setup which has pause between each switch on to allow the magnetic field in the trafo-core to decay. A microcontroller programmed to do this would be the ideal solution. Also, the use of IGBTs designed specifically as automotive ignition coil drivers would make the circuit very resistant to voltage spikes etc. Fairchild has coil drivers in their product portfolio. I am working on a setup like this and will upoad a video if it works out ok..
I enjoy watching you work on these. Consider me a subscriber seems radio guys are a dying breed so always looking for more people to learn from. Also am loving the hot takes on the different audio that comes through.
as a car collector all i can say is i hate it when people put a new radio in an old car, great that you made this video 👍
What appears to have been overlooked is that there was a capacitor in parallel with one winding of the transformer, forming a tuned circuit with the transformer inductance. When both vibrator contacts to the primary opened, the tuned circuit would start a decaying oscillation. After a half cycle, the polarity on the transformer windings has reversed. The vibrator was designed to connect its other contact as close to the voltage peak as practical within tolerances. There is a short time when both contacts are open, just long enough for the capacitor to reverse polarity. The voltage difference across the contacts, just before they closed, was minimal, greatly reducing arcing and extending the life of the contacts. The energy stored in the transformer inductance is not lost. It is temporarily stored in the capacitor and returned to the transformer, but with the transformer's magnetic field reversed.
If the capacitor is removed and the FETs do not provide an interval when both are open, the energy stored in the transformer's inductor will be dissipated in the FETs and converted to heat.
In the vibrator design, If the capacitor were bad or off value, the vibrator contacts would have a short lifetime, due to arcing. Replacing the vibrator would get the radio working again, but the contacts on the new vibrator would burn out quickly and the new vibrator would fail. The vibrator was socketed for simple swap out. If the radio was burning out vibrators, it needed to be removed and checked for failures of the capacitor or other components in the power supply.
Vibrator power supplies in car radios were used until about the mid 1950s, when power transistors for the radio's audio output stage were developed. Then, "space charge" vacuum tubes, which operated from a 12 volt B+ supply, were used for all the radio's circuitry except the audio output stage, which had the only tube which required the higher B+. The expensive, heavy, vibrator power supply was designed out, and the art of designing them was lost. The "space charge" radios were in production for only a few years before being superseded with all transistor radios. However, car radios were produced in high volume, so millions of "space charge" radios were produced before the design was removed from production.
I started cracking up when the guy thought that beer was a health drink.
Well it is ya know. It and Corn Flakes are the breakfast of champions. Lol
Beer is liquid bread!
I'm fairly certain that those vibrators ran at 115 HZ. You don't want ot run much lower as the transformer will draw more current due to core saturation with the lower frequency.
By any chance did you (or do you) have a relative named Billy from Texas?
JAMES AZBELL I am from Ohio and I have brother in Birmingham AL. Iwould use a microprocessor solution to that vibrator replacement, but you have done a great job with analog.
Radio shack still makes the proto boards with a solid bus in the middle. 276-150 is the number.
First one of these I built was about 1967. I'm 75 Back then of course were a pair of germanium power transistors
18:56 . . . The comment about the FETs exploding into fire, instead of damaging the vehicle's electrical system was gold.
I actually did something like this a few years ago on a Motorola radio. here is a list of the things I did.
Instead of using the CD4047. I used a TL494 chip. This is a plain vanilla generic PWM driver chip that can be found in about 90% of the computer power supplies and high powered car amps out there. The nice thing about using this chip is I can now have voltage regulation. The B+ will stay constant regardless of the voltage on the car's electrical system.
To run the vibrator step-up transformer at 12 volts. I increased the frequency from 115Hz to 340Hz. The reason for doing this is to prevent the core of the transformer from saturating. I also replaced the rectifier tube with solid state diodes.
As far as the tube filaments go. I used both a dropping resister and a linear voltage regulator. I used the dropping resister to bring the voltage from 12 volts down to 8 volts. Then I ran the 8 volts into a 7805 linear voltage regulator with the ground pin floated with 2 1n4001 diodes. this gave a regulated output voltage of 6.2 volts. The reason for pre-dropping the voltage with the dropping resister was so the VR didn't have to brunt the full voltage. the VR only had to handle a 2-3 volt drop instead 6-8 volts and therefore, runs cooler.
Thanks for the updated info.
Just put a lm317 voltage regulator on your board and adjust it to proper voltage out from the transformer. That lets you ditch the diodes and the zener’s.
Built a lab supply using them back in college 30+ years ago. They're stable...I give you that.
solid-state vibrators are sold on the website of antique electronic supply (tubes and more)
..and they are noisy as hell..
@@blitzroehre1807 solid state noisy? rf interference on radio?
@@bask2140 Yes, RF hash from switching harmonics, some designs have a lot, some are pretty good.
Very Sad to hear about the "This is a test" girl". I will miss her. I hope the Zombies didn't get her ? Excellent Video , We use to take the old vibrators apart and gently sand the contact points. We would then put a tiny drop of Transmission or power steering fluid on those points. I had them last as long as 5 more years. ( our old radio was used daily ) ( 1954 Ford radio)
The old 0Z4 tubes were in just about EVERY GM radio back then. The old Fords ( Bendix or Zenith ) used 6X5 rectifiers. The tubes were cheap & plentiful. Few folks fix these old radios these days. They just don't even bother teaching this old technology in school today. Thank You for yet ANOTHER Excellent Video.
Using CD4000 logic really goes well with the general theme of the show. 👏
Hi thanks for the video, the radio has a nice system for tuning , it modify the inductance instead of capacitance.
Product design versus electrical needs is always the toughest nut to crack. The offerings by sellers of solid state vibrator
replacements offer no guarantees. Makes sense after reading Paul Carlson's comment.
12:15 "This is why I don't like working on stuff in the house"
Agreed. Especially when you notice specks of Dust moving.
Because you have just introduced House Dust Mites into your home.
103 AT 2KV
youre a poet and didnt know it.
What I find impressive (besides how well this ended up working) is that all that was needed to replace a mechanical vibrator. Score one for old school!
Now that's a good idea, to use a PDM (or PWM), you could make it into a full fledge regulated sw mode power supply inverter!
I love how he turns the lamp up so we can see the bulbs, kinda reminds me of the RCA dog or the Luxo lamp from Pixar. 🙂
RIP 'This is a Test, This is Only a Test'.....cheers.
Very nicely done would love to see a video of it in the truck later on
The way I would do it is rewiring the transformer, that is, just use two IR2153 as a full bridge driver and connect the whole primary across it. Heater wise, make a single transistor running at low duty cycle, switching the heaters on and off from 12V.
So fascinating! I like that resistor. Imagine if you could use the heat coming off of it for something.
No need to add that masive resistor to drop B+, you can add a series variable resistor to that 22k on pin 2 to adjust the frequency. Also if both fets get hot it could be because the output of the 4047 might be turning both fet at the same time similar to the crowbar state on a H-bridge of a switchmode power supply. You can try adding a delay on one of the fets or use an oscilloscope to see how they are switching. The gate of the FET is a capacitor which means that if the output of the 4047 is not pulling it to ground fast enough it will create this problem. Or you could change one of the FETs gate resistor to a lower value. That way you'll be adding a delay.
The minimum frequency of that chip is 100 Hertz so what am I going to do speed it up to the point where it's out of the Transformers response? I can see adjusting the duty cycle yes not the frequency
@@shango066 are you sure?? Even if you change the cap?
@@shango066 Since the goal is to reduce the output voltage, tuning it to a less efficient point in the transformer's response curve may be a reasonable way to accomplish that as long as the transformer core doesn't get too warm. Alternately, 555 is the classic way to do a variable duty cycle low freq square wave, and it's easy to make an inverter to get the other phase gate drive using a transistor and a resistor.
@@shango066 I dont have a 4047 to test, but I did a simulation using proteus and changed the timing cap for a 220nF and a 22k resistor and I have a frequency of 42Hz
Monostable multivibrator circuits are also useful in creating a nice sounding click from the weak pulse output of some vintage geiger counters like Eberline and the Civil Defense CDV700. You can find some schematics online if interested. Their main benefit is super low battery consumption so one battery can last for years if usage is occasional and you remember to turn the power off.
Way above my skillset but very interesting workaround.
You and Bob Anderson should do some sort of vintage electronic event. Add Mr Carlson, radioTVphonorant, Jordan Pier, MrRadio, Buzz1151 and Glasslinger.
And despite his being in Europe, I'd love to see more from TRX Bench Peter.
The diodes are dropping the voltage to the oscillator chip and hence dropping the gate drive to the fets. If you run the oscillator chip on the full 6 volts to maintain full gate drive, all should be well. Good luck with the project and thank you for your entertaining and education videos over the years. On reflecion, it would make more sense for the diodes or resistors to connect to the centre tap of the primary and leave the drive circuitry at full power.
That's one reason why I like the idea that Jernej Jakob / braindamage had, of putting the dropper diodes in the two legs of the transformer primary, rather than on the center tap. That keeps up the voltage to the chip, and the FET gates. One downside to that method, though, is that when the primary rings, instead of the back-voltage going harmlessly through the built-in snubber diodes in the two FETs, the series dropper diodes will block the backflow, which could lead to those series diodes seeing some hefty reverse spikes (depending on how fast the FETs are switching on). Simple series dropping resistors in the two primary legs (instead of diodes) could avoid those problems. The thing is, since Dan doesn't want the RF noise of a boost-buck converter, you're going to be dropping voltage somewhere, with diodes or resistors, and that means heat, so he has to choose where to locate those components so that they don't cook the little FET board.
@@auchterawer1150 Further thoughts: I can't see that the FET drive will produce any more B+ than the origial vibrator with it's metal contacts. 350V may be a bit on the high side but if it doesn't exceed the tube or capacitor ratings why worry?
Well done Sir!
I guess you can't adjust the Mark to Space or Duty Cycle as with a 555?
It seems as though this circuit would really benefit from having a Set H.T. Preset.
i added a 4.7uf on pin 8 to ground to make it soft start if that feature is wanted on the design
I think the heating encountered when putting the diodes between the supply and the center tap is caused by the inductance of the transformer. When one side turns off the current wants to continue flowing so the voltage shoots up. Normally the current continues by flowing backwards through the other winding, but with the diode present it can't. So the voltage shoots up as the transistor turns off and causes excess heat.
Putting the diodes where they suggested would lift the source voltage above ground effectively lower the difference between gate and source, the fets would probably have been running in their linear region
I really appreciated and enjoyed this video on building the solid state vibrator circuit, great job there Shango066. I have a small collection of vintage tube type auto radios. So far, I've been successful opening the vibrator "can" and cleaning up the contact points, and reassembling them to work. Someday I will build one of these SS vibrators and try it out! On another subject...I wonder what happened to your "this is a test, this is only a test" radio station? I think you mentioned before that it had been on a long time. I wonder who was broadcasting that signal and who was paying for it.
Good progress. Maybe you could add some resistance at the solid state rectifier? Looking forward to the follow up.
It does not like 2 diodes because you are lowering the voltage of the CD40XX chip also and it needs at least 5 volt or more to work correctly and with an input of 6 volts and 2 diodes in series with the circuit the CD40XX chip is not getting the required 5 volts to work. You should change the R/C components around the chip to obtain the desired output voltage instead of adding diods to drop the voltage. Or you could use the NE555 8 pin chip to do the PWM for the mosfets and it would be totally able to regulate the voltage the way you want by using a little pot to do that.
Fet's have a gate threashold voltage, commonly 4v, if the power supply drops to around this voltage the fets will go into linear mode due to low drive voltage, and will get hot..
Bipolar trannies may have been better at 6v power.
I'd like to try this with a royer oscillator, the o/p is close to a sine wave, and the circuit is simpler.
That .004 capacitor on the secondary is probably to reduce ringing of the transformer, you see caps like that across the secondary in many tube radios with solid state rectifiers to reduce ringing caused by the slow and abrupt recovery of solid state diodes. that cold cathode rectifier tube may have had a similar probably so they put that capacitor there.
Revolutionary
Vibrator? Oh, hell I'll behave! Cool video!
Could you use a 4-6v regulator instead of the large resistor? Regulate the voltage down to 4v and you may not need the diodes.
Sg3525 would have been a good option you would need a boost converter/charge pump or logic level fets but ive got a feeling the sg3525 will want 12-15v it has UVLO and soft start to protect the fets it's a great ic fully adjustable frequency duty cycle dead time etc also a powerful fet output so you can drive fets directly, it virtually a tl494 but with a nice fet output stage , it has 2 error amps so easy regulation.. its not pin compatible with the tl494 however.
I think diodes are not an intuitive application for voltage breaking? Seems like that is a job for resisters and / or a potentiometer.
With a radio that needs 220V, anyone who doesn't want to do much soldering and doesn't mind having something conspicuously 21st Century under the hood could just order a 12-to-230V power inverter intended for a high voltage country like the UK or Australia. I see a bunch of them on Google from $20 to $500, probably depending on how much Chinesium you can tolerate.
I would love to see the performance of that radio far out in a desert or cave location seems like a really hot set
if you used an H bridge across both windings of the transformer at once (instead of the push-pull) it should be possible to run it on 12V since you're using both windings at once instead. but of course that means an H bridge driver instead of just two fets. I don't think the transformer would like 12V across each winding even when PWMing it since it would cause it to saturate.
Wouldn't you have to separate the two primaries to do that? And then too...wouldn't firing them simultaneously 'buck' each other and cancel each other out?
@@Rev22-21 the primary is center tapped, so the two coils are actually out of phase with each other. this is how it works as a push-pull. If both windings were in phase it wouldn't push-pull any more. By not using the center tap and leaving it free floating, this converts it effectively into a single winding instead. You need to drive it with an H bridge though, because doing this loses the benefit of having the center tap to reference from. It's similar to how you can make a full wave rectifier with 2 diodes and a center tapped transformer, or use a single winding and a 4 diode bridge.
By the way, it appears that those black capacitors at 14:00 *_are_* already marked as to the inside/outside of the foil with the little nub by one pin...
To drop the 12v to 6v for the radio, maybe a 6.2v high wattage zener diode in series with the power lead to the radio would work better?
Actually, using the ''3525 chip and replacing the tubes with their 12 volt versions (14xxx series) might be the best idea. The transformer is a 6v one, but you could probably run it on 12v by cranking down the duty cycle and putting diodes in series with the power input to the inverter. Then you wouldn't need that big ass resistor! (Yeah also have to replace the pilot lamps too).
Back in the '70s, a friend of mine built a DC-DC inverter for a ham mobile rig using a power transformer with a 12vct filament winding being feed by a pair of cross coupled 2N3055 transistors. The filament windings were used as a push pull primary, and feedback was crude, power resistors from the base of one transistor to the collector of the other, capacitors in parallel with the resistors. We played around with the values to get the best efficiency and output voltage. The HV secondary was used as normal, the primary was used to generate bias voltage for final amp tubes in the rig. The rectifier filament winding wasn't used for anything (NC). The thing worked, but it was sorta Rube Goldberg.
When I was installing a 12 V negative ground stereo in my 12 V positive ground Triumph, I had an inverter-like device I got years ago. It could also be wired to produce 12 v from 6 volts and vice versa depending on how it was wired. But I haven't seen anything like it lately. It came from Radio Shack. Probably would have been simpler just to convert the car to negative ground.
A very nice solution.
Always wondered why these old car radios always seemed to use slug tuning rather than a cheese slicer. Is there a good reason why?
Vibration. Those variable cap plates would vibrate and change the tuning. Slug tuning is much more resistant to vibration.
@@eeengineer8851 Ah, never considered vibration. I have a Granco FM-only table radio that's slug tuned which performs very well. I never figured out why they went with that instead of a variable capacitor either.
@@defaultuserid1559 tuning capacitor for only 88 - 108 MHz is expensive for what it is. TVs were using slug tuning and then went to a cheaper form capacitor tuning for a while before they went to varicap diode.
I have seen rectifier tube deletes using modern Silicon rectifiers using series resistors to lower the B+ Voltage. This would lower the regulation but no more than the original transconductace value of the tube. The current drain should be "constant" providing the audio output is class A. The AGC may affect it slightly though.
Normal running voltage should be about 7 to 7.2 volts. Of course old generators put out nothing at idle so you could go down to 6 or even a little lower with lights and heater on. Car radio tubes were supposedly designed to operate with the varying voltages of old fashioned poorly regulated generator systems, although I think any indirectly heated tube is pretty tolerant of varying heater voltage.
I wonder if you could create a voltage divider with a h4 bulb connecting 0v and 12v to the outsides and the common and one outside for the radio
Well done on the vibrator mod. You could try replacing the 22k timing resistor with a pot (say 100k) and use that to change the oscillator instead of using diodes,' but, needless to say, test out of circuit (with test transformer) first.
I need to replace my vibrator. Is it worth getting the replacement or is the solid state a better option? Would you make me one if I paid you? I’ve replaced all the tubes with no luck.
wtf no more this is a test
my favorite radio station
I wonder if you could get away with a ZVS and ferrite transformer, the output on the secondary is sinusoidal, run the gate supply directly from the 12-14v and use a BJT pass element on the input to the fets to regulate the output B+ on the secondary, I think you might be able to get away with it with a bit of filtering etc, it would preserve the original transformer too
needs short circuit protection. Current sense resistor and a transistor to disable the MV. No need for the 1 00000uf crapacitor there. remember the Royer circuit uses a impedance, an inductor in the B+. The Diodes in the source reduces the drive too much already with 6volt
"More Dr's smoke Camels" 🙂
Risking a repeat here (of what I asked of Mr. Carlson) has anyone considered using a PIC in place of the IC? An 8 pin PIC could be easily programed to control those FET's. A few lines, a little time and presto....
I would have no clue how to program it myself.
@@shango066 : The following YT Channel does a pretty good job at explaining the A-Z's on PIC's . The Channel is simply called Learning Microcontrollers. I've built two major projects myself; one (a one year long project) was patented and one was recognized & published in Coca-Cola USA Tech magazine. The former started out as a logic IC topology, went through seven board upgrades and finally into the PIC Microcontroller arena. I never looked back.
@@shango066 Learning how to program is like learning how to play a musical instrument, rebuild a carburetor or put a puzzle together. You watch others "program", repeat what they do and before you know it you're making an LED blink, making a switch turn on/off something, building a timer of sorts, monitoring sensors and then building thermonuclear war head (not really). But you get the idea. 😜
It's like moving from discrete components to IC's. Simple.
Maybe try two rectifier diodes in series then add a cap to ground at the centre tap to average out the ripple voltage, it might bring it in to the sweet spot and also take some stress off the diodes?
By putting diodes in the source you raise the source to 1.2V, and then probably the MOSFET is not opening full = heat. I think you had that figured out the first build. You measured the gate as it was a bit above 2V, and that will stay. Just put a few diodes in the positive, as (i think) you did the first time ? You can put as many as you need.
You did at the end...
fets need heatsinks too and if you are running them long term without heatsinks they will burn out if they overheat.
Why not a 12v-6v voltage regulator or a Zener Diode?
I would've tried a DC-DC converter to go from 12V to 6V. More efficient and less heat.
The noise generated by those things would make the radio unusable
Or maybe instead of a resistor and a diode, use a varistor, which will also play an additional role of protecting the vibrator circuit?
30:46 how mild, how mild? how mild can a cigarette be....try the camels 30 day test, and you'll see. TRY CAMELS AND SEE! 🤣
Almost did a spit take when I heard my voice 😂
nice capacitor checker
Very cool indeed.
cant you change the freqenty of the vibrator board to lower te output voltage? in that case yo dont need the diodes.. (its a pwm chip, so mounting a potentiometer on it would make it adjustable for the input voltage of 7 volts.whitout extra heat...)
What about rebuilding that set using the tubes designed in 1957 that has a H+ of 12V and a B+ of 12V and using the two VMOS FETS as Transformerless Class A output stage to drive the speaker.
24:00 pot it for life
in 50 years, Mikes radio's son can dremel it out like a cybernet pll VCO can.
why you removed the video..??Is there any design flaws??
I am I am I said I'm not myself
I'm not dead and I'm not for sale
Hold me closer, closer let me go
Let me be just let me be
@@shango066 Awesome....
Well, that would be great indie song lyrics Shango. I will put you in Written by category if I ever end up dping this song.
@shango066 Trippin' on a hole in a paper heart. If the trip was caused by a wet floor contaminated with Camp Lejeune water you might be entitled to compensation.
I read about you are not able to se a DC-DC .... sure, those are noise sources. You can use an LM350 mounted directly on the case, (use the whole thing as a heatsink). Not the pretiest .... Not sure if the 350 will like the load at all. It is a shame because you need to modify the radio a bit. At least you can get some control. Be sure not to starve the IC or be to much low that Vgs on the mosfets are not high enough so they wont turn into heat sources.
Do a part 3 Shango 👍
Fantastic content
Imagine trying to replace a Dynamotor power supply for the OLD MOTORONI police radios! NOT! Can you imagine what the factory charged for the "radio option" in the day?
Surely a buck regulator would be a better plan to convert from 12v to 6v?
Yeah, resistor drop will depend on current. Maybe it is a divider, but that would kill the battery in no time.
noise
noise
@@shango066 Hmm.. they can be noisy, but sticking it in a metal can and filtering the input and output would probably sort that out. Might be worth a try if you have issues with that space heater of a resistor - would give better DC regulation.😀
Is that schematic and parts list available?
could you use a 6 volt regulator? Triumph 6 Volt Regulator
if its ac at the end, why not to use just a small transformer
3:05 I'm disappointed! I was hoping to see you modify the radio with 12-V-plate space charge tubes then build a one-transistor output stage with a giant heat sink. It's important to keep up with the times, you know.
Ye Gods ! Electronics from the Middle Ages! Talk about crude and primitive. To get the final output voltage regulated don't have series diodes and dropper resistors as you'll be faffing with them forever. Add a voltage control loop instead. Pot-down a sample of the B+ output voltage and when it goes over a set threshold turn off the FET drive. Can use a Vbe junction as a comparator. The circuit then runs in squeg mode. Does the B+ voltage really need regulation anyway? Valves (tubes) and good capacitors can tolerate huge variations in anode voltage. The 12V supply input ranges from 11.5V to 14.8V anyway. EDIT: the second schematic with the generic SMPS IC , it's being used here 'open loop', normally is used closed loop as I describe above. This chip has a built in switching crossover function to prevent both FET conduct at same time, I think it's 2% though it's been 30 yrs since last used this chip.
why not to replace the vibrator and the rectifier tube and use an external 12 to 6 v converter? As a radio collector I do not like this solution , I like the video anyway