TSP #56 - Teardown & Repair of a Stanford Research PS350 5000V, 25W High Voltage Power Supply
Вставка
- Опубліковано 29 вер 2024
- In this episode Shahriar repairs a Stanford Research Systems Model PS350 5000V-25W High Voltage Power Supply. The unit continuously displays 2.5kV without the output being enabled and produces no output voltage. Verification of power supply voltages reveals the issue is linked to a disconnected 15V voltage regulator IC. After the repair, the output voltage is verified with both positive and negative outputs. The principle operation of the instrument as well as the Cockroft-Walton high voltage generator is reviewed.
The Signal Path
www.TheSignalPa... - Наука та технологія
This is exactly the keithley (or vice-versa) 248 5kV power supply. Interesting... As always a nice video from Signal Path!!
For an improvised HV divider I'd be inclined to use the uA range on the DMM - the lower input impedance should reduce the risk of damage in case of an arc-over.
Even with a resistively limited circuit, a spark from a connection coming loose can cause issues due to the very fast risetime combined with cable inductance.
mikeselectricstuff Thanks Mike, I didn't go over 1000V, beyond that I used the output monitor voltage. I'll make sure to use the uA range to try this out as well.
Repair episodes are very interesting and informative. Keep them coming .
HI Shahriar, great video as always. Only one very important note for safety that you have not followed I'm afraid...
---> in ANY non working HV equipment, ALWAYS work with one hand in your pocket and one hand on the device (you should have clamped the ground). Electricity always takes the shortest path to ground, and if its unavoidable to have that path go through your body, you might as well have it go through in a way that won't kill you. Not having a conductive path hand to hand is important, will never go through your heart or other vital organs, instead it may find ground through your arm and down your leg, saving your life.
Just my 2 cents!
Kris TheTrader Electricity doesn't always take the shortest path to ground. It takes ALL paths to the return conductor.
another very interesting video, I never seem to get the kind of easy fixes you do! Looking forward to some high voltage video!
A nice permanent fix would be to extend the leads with some stranded wire.
This happens very often, even without added mechanical stress just the thermal expansion can break solder joints if the heatsink mechanics aren't done right.
stefantrethan Even worse in the days before THP and pads were very easily ripped from boards.
I can't wait! Watching it at 360p.
***** ha! Same. There wasn't even a thumbnail when I started watching.
Such a nice instrument, let down by the lack of an interlock to prevent the polarity being reversed while powered on. It would have been simple to implement, and hardly a significant added cost given that the psu must have cost a small fortune. Repair money for the manufacturer from blown units maybe trumped the desire to design it better.
My PS325 cannot go to zero (abt 10 volt residual) either, with an differential probe, i get an extremely narrow PWM, which makes sense. I use it for PMT in gamma spectrometer, very handy.
nuo xu looking at the PS300 series manual, the output voltage for the PS325 is specified as 25 to 2500V, and the PS350 as 50 to 5000V. So the designers apparently knew the voltage divider noise at low output voltages was a limitation, but didn't see it as important for the typical usage of these units.
I've been wondering.. is there a specific reason your intro is telecined?
Btw, just discovered, Keithley did sell this unit under own brand, as Keithley Model 248 (www.keithley.nl/products/dcac/highspeedpower/highvoltage/?mn=248E)
That's almost cheating, to have such an easy repairs!
Maybe you can get some broken old Keithley gear, which is always PITA to fix instead :D
Anyway, congratz, enjoying higher F of production too!
P.S. Would be interesting to supply analog signal from soundcard playing some mp3 and get arc sing for you :)
Don't those linear regulators have to have a load to be at zero?
I always thought Zilog was pronounced the first way you said it until I found out otherwise...it's pronounced 'Z-log'
Very interesting failure with excellent example for bad mechanical design..
Nice repair! And a great lesson for chassis designers, as you said. Thanks for your videos.
The bigger the glob the better the job.
Applied Science Thanks, to you as well.
Instablaster...
The most imoprtant and enjoyable part of these videos is the excellent scientific method verbalised at at every stage of all repairs
Kitty! Everyone needs a lab cat!
Connor Wolf Except during HV experiments :P
Connor Wolf why don't you make videos anymore, connor?
Connor Wolf glad too see you're ok :)
This episode was also extremely funny! Hahaha
Another thing I don't know if you noted is that the chassis seemed warped when you had it on the table (one of the LRFs (Little Rubber Feet)) was in the air (dunno if you have a warped table).
Could be that the unit has warped over time or got dropped, which in turn caused the pin failure of the 7815.
Any ideas on trying to secure up the mains transformer so it won't affect the chassis as much?
If this could be done safely and enough power is available. Ramping the voltage up on a magnetron till it just started to oscillate and looking at the wave forms.
You could do some experiments with vacuum tubes perhaps.. 300-600v would be good for normal power tubes, and then some of the big transmission tubes can get into the thousands of volts, although I'm not sure at what amperage.
For experiments that look interesting, you could mess with some eye tubes, NOS ones of certain types are quite inexpensive. Tubes are quite hard to burn out so it could be fun as a kind of retro blast from the past.
If changing the polarity while the power is on would damage the unit, I wonder why they didn't just use a 3-position switch with the centre position "off"?
thelastofthemartians
even a two-position switch has an off position in between, you just can't lock to it. But I doubt that's the problem anyway. The problem will be arcs when switching that burn/oxidize the contacts in no time. An off position won't help with that.
Wow how do you keep getting lucky with these easy to fix problems. I need to start buying cheap "non-working" shit on ebay and start doing repairs.
***** I was thinking that too, but say have a patreon 'raffle' / lottery, else he might run out of storage space :D
***** Without Shahriar's luck you'd get a broken HV transformer or another irreplaceable part
Don't apologize for the frequency of the repair videos. We LOVE the repair videos!
You might want to keep an eye out for a AN/USM-223 (aka ME-297/U) on the used market. These are all-analog VOM'S that include 1 and 5kV ranges. They're remarkably accurate, given the technology with which they're implemented. They're ex-military, so they score 10 on quality and 0 aesthetics. When they first hit the used maket they were as little as $5; by time I found out about them I paid $25. They still show up from time to time; you be the judge on price. It probably would have been useful for testing this power supply.
you should buy pair 464-IEC-RT or PRUEF2610FT RT probes they are absolutely brilliant for measurements on small pads.
12:47 Gear teardown rule number 1 - if you see a "BIG CLUNKING POWER SWITCH" in a teardown, you have to press it. Its the law.
i have a problem with a dc power supply; is not addressed on you tub. the dc supply suppose to source out 20 -220 and 130 - 220 dc volts but can only read around 50 to start with but then it keeps dropping and stops at 30 vdc,,, why???...is like a capacitor discharging ...any ideas fire them now or don't pretend to be a scientists by making 10 years long boring videos
I would go one step further in your analysis: transformers naturally vibrate at 50/60Hz, so you even have a permanent stress on the pins when it's on. Then the pins being copper, they tend to work harden, so they become more and more brittle with more vibration periods.
Great video as usual. For the voltage divider, i'd have used two or three of those Glass resistors in series just in case there was a manufacturing defect with one of them. I know they are pulse rated to 10kV but you can never be too careful.
Hi. We have a PS325 model. And the Voltage output only displays 0 even when we set a voltage value. When we turn on High Voltage, the "Limit" at status display also lights up but out voltage display is still +0 and current display is also 0mA. Do we need a load to make the voltage output to display the set value?
Would it be possible to setup something like Repair Friday.. like a dedicated day of the week for repair videos?
I really love this repair videos and I am sure many other views do too... so having 1 of them each week would be really really nice!!
Hi. I saw this bad chassis design in many devices. Make us say "what engineer was thinking?". You got a new subscriber.
this device has obvously been bashed around so those pins probably broke in 1 big event because of the transformer weight. It might be worth resoldering the other regulators too.
it's not unusual that a DC to DC converter has a minimum output voltage or output power, especially with such a high plant gain. i wouldn't expect it to go down to 0 like a linear regulator might.
Shahriar, put a metal bracket between the HV can and the side wall or between the two sides. That will stiffen it sufficiently to prevent this sort of failure from happening again. Really crappy mechanical design, indeed - someone wasn't thinking when they put the heavy transformer on the side wall.
The other option is to desolder the regulators from the board, turn them upside down and use wires to reconnect them. That will provide for a bit of give and prevent them from shearing off.
It looked as though the chassis was warped in the beginning of the video. It is possible the weight of the transformer caused the regulator pins to shear as a direct result of the laws of gravity :D
M@ You mean when it hit the floor:-)
A simple 3cm 'L' made to the PCB either side of the V-regs and screwed to the chassis solves the problem of further failures.
Shahriar, how do you connect that wire and get a good contact from the middle of the female SHV? I want to do this; I will not go any higher than 800V.
Great video! You could try building a Ionocraft en.wikipedia.org/wiki/Ionocraft
Could you inject audio into the voltage input to make the unit drive electrostatic speakers? I'm thinking there will probably be a frequency limitation due to the capacitors in the output circuit.
MichaelKingsfordGray
I didn't realise flame speakers existed. I might have to give that one a go one day. I've seen plasma speakers where a high-voltage flyback is employed to produce a constant arc that is then modulated.
Judging by how slowly the output voltage responds when the setpoint is changed on the front panel, I'm guessing it will not pass audio.
If the GND pin of the regulator fails 1st, you may damage other stuff ;)
the legs either vibrated apart or I was thinking maybe someone wanted a new instrument and just cut the legs.
ha, that actually might been the case
Good idea to clean the HV section with alcohol, this voltage does attract small conductive dust.
It doesn't give zero when you want zero because it's a 5000v supply. Why people will use a 5000v supply with 2v, 0v?
Navarro Eletrônica I explained the likely reason why it does not produce low-voltages well. Although the main intended use is not at 0V, it is important to know why it behaves the way it does.
The Signal Path Blog I got it!
I'm working to repair a used PS325 that I picked up on eBay. Unfortunately, someone rather incompetently tried to repair the unit before it got to me (including using incorrect replacement part in some places.) So, I've had to work carefully to undo this bodged repair and put back correct components. However, this is made harder by the fact that the manuals available on-line all seem to lack schematics and are either incorrect for the version of unit I have, or just plain wrong in some places places. If anyone seems to know where I can find a printed manual with schematics, I'd appreciate it. The PCB in my PS325 is labelled as part # 7-00184-701. REV E, 1989
Turns out Stanford Research sells printed copied of the manual for my PS325 that include schematics and board layouts. Cost is $100. My should show up in a few days and, hopefully, will let me get this sucker properly repaired.
this is rated 50V - 5kV, so probably not meant to go down to 0
What's going on? Previously ebay had three "condition" categories: new, used and for repair; now it has: new, used and for shahriar
Andrés Domínguez Haha... :)
In user manual it is written that this unit works from 50V, so you don't have to be worry about the low voltage indications.
Jacek Pawłowski Great, thanks.
What would be a typical application for such an instrument, lab wise?
A. R. Jasso Powering a photomultiplier tube.
I need a lab cat! :)
Considering the price Stanford would likely have charged for this unit new, I feel the poor mechanical design is inexcusable.
Nice repair!
Also, I would pronounce Zilog like Zeyelog (with the I sounding like the I in rice)
***** I bet the cat ran a mile when Shahriar sparked the gap. PS. It's pronounced zilog as in "ill" here.
IanScottJohnston He is locked out of the room when live voltages are present. :)
When I saw the output stage I immediately thought of CRT's. They use the same principle of voltage multiplication, and with flyback transformers. Triplers were commonly used in color television to provide the high voltage for the cathode ray tube (picture tube). As in your case the voltage is quadrupled for that instrument. Well lets hope your next repair is a bit longer, I know it's not your fault, it's a fault of the fault. Great video look forward to the next one, Cheers.
I always enjoy and learn from your videos. I don't have much to say because you do such a good job of fully explaining everything. Thumbs up.
The diagnosis required only minimal electronics a couple of quick probes and some careful observation. This one is a MOTO. Master of the obvious Well done
This was a very enjoyably project, I have repaired many devices with this style of component mounting for heat disposal and many have had this type of mechanical failure. Very common on rental devices that are shipped frequently!
That's a SHV connector. TE and Pasternack make them.
👍👍
I have tried the Cockcroft-Walton on a breadboard to generate 5 kV long time ago. Was able to sustain an arc indefinitely.
GPIB IC sockets layout looks like it may be a solution with NAT9914, 75160 and 75162, so in theory (if the firmware is the same across all units) you could add GPIB with these few chips and GPIP connector over ribbon cable.
Robert Lukierski That would be interesting. I found that OPT 01 on this unit is actually the GPIB which is missing. So someone has actually removed it! :(
The Signal Path Blog It's just a gut feeling, double check if NAT9914 makes sense there (DIP variant) :) I have SR850 so I can check tomorrow what they've used there (maybe they reuse some designs).
The Signal Path Blog actually that circle next to OPT 01 would be filled in if the option was factory installed. At least that how HP & Keithley mark options on their back panels.
For your voltage divider did you take into account the input impedance of the multi-meter when you said it was out a bit? That would explain why the direct connection up to 1KV tests showed it was spot-on.
You did make mention of the input impedance when explaining why there was current draw with only the meter but this would also apply to the resistive divider as well!
Really like your short repair videos (as well as the longer reviews) as I go into "Signal Path" withdrawals when there's a long interval between them!!! And this one didn't even require a little liquid libation to effect the repair! ;-)
How much $ Shahriar?
What a great video repair, and great lesson for chassis "Architects". Very nice and useful Video.
Repair vids are so good
Amazing! You’re awesome 👏🏿. Thank you so much!
Thanks for another great repair video. You make it look easy but that's because you're very good at thinking logically through the layout even without a schematic.
Thou shalt test voltages. :-)
I'd like to see surge protection experiments and testing.
First time I've seen that failure mode. Pretty sad. I wonder how many dead instruments are out there because of this.
Well done. Analysis was spot on at every stage. Even your incredulousness at the poor build quality was totally in order. A+
I can't recall seeing a failed repair video in the past, but if you do fail - it might be educational to see that as well. Wonderful video as always
600-th thumb up
With such a low current, will 5000 volts really hurt someone?
Razor2048 While 5 mA is generally not enough to kill, it's likely quite unpleasant. Also, the peak current available might be higher due to the filter capacitor on the output of the multiplier!
Razor2048 5kV@5mA = 25W. Yes that will hurt!
I like the fuse holder cover that makes you to unplug the AC power
good job.
first time see high voltage PSU.
Thanks a million Shahriar, helped me huge
nice video, thanks.
Thanks for the video Shahriar
What do you mean by "I downloaded the firmware from the IC" on 33:35? Are there specialized tools for this?
MrKayPham An EEPROM reader.
pls more repair vids !!
KITTY!!
Amy Marie KITTY!!
Tesla coil input?
joblessalex That would require an AC source, this is a DC power supply. :)
The Signal Path Blog Indeed, but I was thinking a high voltage switch of some sort on the output that you could even modulate for sound.
love the videos!
In 30.40 minute you say: ".. so I not know what is output impedance of that thing."..
Well I have problem with that, in DC you do not deal with impedance. But I would be very pleased if you correct me with it.
Other than that ,,,, wow... I love your videos.
Wesley
stivep1 Hi Wesley. You always deal with impedance at any frequency. Except that the reactive part of the impedance would be zero. The resistive part is non-zero. You can define the output impedance as a function of frequency. When frequency goes to zero (DC) the reactive part collapses to zero but the resistive part remains.
The Signal Path Blog Wikipedia:"When a circuit is driven with direct current (DC), there is no distinction between impedance and resistance; the latter can be thought of as impedance with zero".
I have the same PS350 5000V in my lab. I did not check it with oscilloscope for ripple. That would be interesting. but with over 400 toys all around I simply did not think about it. So my point is that at the output we have DC, and by that only reminding component is pure resistance. The reactive part is not present.
That is why I questioned your "impedance output comment"
For me you are clear in most of other material presented to the point of satisfaction. Thank you for your work.
Wesley
stivep1 Thanks, the Wiki statement is the same as what I am saying as well.
Do you have a HV cable for your PS350?
The Signal Path Blog any HV cable meets the needs.Look at the ebay section of "particle" spectroscopy. Every HPGe or NaI(Tl) must have use of HVDC.No exemption.( however HV divider is often build into socket of PM.
Plenty of junk for $40 or less for Beta co/anti-coincidence spectroscopy you get from Radiomatic Fow one .search for (CANBERRA RADIOMATIC RADIO CHROMATOGRAPHY DETECTOR) You get beauty in it. 2 nice HV BNC
and you can run "repair on it" wow that would be something.... This thing usually fail with on board batteries. I have 6 of them with screen. ( model A525). do not be discouraged by price you can get it below $100 or go to auctions usually $25 per piece. I would describe it as top excitement and addition to avenues you are at. It is like fresh juice to your channel. If you have problem with soft for comp. let me know.. (no charge) I'm in spectroscopy big time and I got the best of its class.(some of them: Canberra-Lynx, InSpector-1000, Detective-Ex, EDXRF SEA5120 +++ )
You can also buy cheaply electron microscope and vacuum chamber. down to (10 to -8 of millitorr) One in the life time unique moments you can have just for few bucks, I got it all.
Wesley
stivep1
Impedance includes resistance, so even at constant voltage you deal with impedance. Other than that DC and constant voltage are not the same thing! This device can be modulated quite easily. As soon as you have changing voltages/currents you need to look at impedance. You don't need the voltage to change sign for reactance to come into effect.