Yes, it was great to see the reverse engineering and the way you could build the schematic back. More of these skills and how to achieve this level of competence will be of great value. Thanks a lot!!
Wow.....what a fantastic and professional review, analysis and explication! Many thanks. I have a few of these in various guises but haven't got around to using them on a project yet. At least now you have got me more up to speed on them. Thanks again!
Fivefishdiy asked if there is short circuit protection. Although the LM2569 has thermal shutdown and current limiting protection I found in practice that shorting out the output tends to damage the LM2596 and it dies. I have been looking at ways of adding better true short circuit protection in a future project and plan to use the Linear Technology chip LTC1477 which has built-in short-circuit protection up to 2 amps.
+Scullcom Hobby Electronics I think it may well happen that cheap Chinese regulators use fake lm2596 chips that are not in par with branded chips. At least I saw fake power transistors on youtube and other fake ICs like lm7805. So I never buy ICs from China.
Having had a closer inspection at the resistor values of the potential dividers at the input and output to measure the input and output voltage the resistors I suggested were 300 ohm are in fact 2Kohm 1% resistors. It seems that these resistors are marked as 30B (using the EIA-96 coding for surface mount resistors having a tolerance of 1%). All the other resistors on the module use standard 3 digit coding system.
Great review, thanks for taking the time to draw and analyze the schematics. One puzzling thing though: the 5V reg XC6203 seems to be rated at a max input voltage of 12V only, according to the datasheet.
LANE Boys RC Yes you are correct. I did wonder myself if the XC6203 would be damaged at higher input voltages but so far it seems to be O.K. If could be that there are different voltage rating types of the same IC and I will investigate this further.
Best review of this sort of device that I've seen. Especially your reverse engineering, very nice schematic and explanation of the unit's operation. Just a few comments. (1) You mentioned several times that it meets its specifications for output voltage, but in fact it never did. The closest I saw in the video was +110mV as measured by your Fluke DMM, and most of the time is was +150mV or more. Since your Fluke was sampling at the input to your DC load, it was actually even worse, if you account for the voltage drop in the wires from the Drok to the load. This should be measured at the Drok board output. (2) Adding a heatsink to the top of the LM2596 (as is commonly advised by the sellers of these devices) will actually do very little. The plastic case has very high thermal resistance and is not intended to dissipate heat. Surface mount power devices like this one use the metal tab on the bottom, soldered to the PCB, for that purpose. A heatsink stuck to the bottom of the PCB would be more effective. See page 25 of the Texas Instruments LM2596 data sheet. (3) The low voltage behavior of the device was pretty shocking. If the display is on, the output should match the display, and it didn't. The ST uC should be sampling the input voltage and holding the LM2596 enable low with the display off until the minimum dropout is met and the device can produce the displayed output. Only then should it switch on and show a value. (4) When measuring ripple on a switching converter, the high frequency components of the signal can make the measurements look a lot worse than they are, particularly if probing is done incorrectly. You really shouldn't use the croc-clip ground lead on the scope probe for anything above a few kHz. The probes come with a little spring that fits over the ground barrel if you remove the spring-loaded hook. The tip of that spring goes to ground and the probe tip to the node you're measuring. This greatly reduces inductive noise in your measurements. / Anyway, thanks for a very nice video. Cheers!
FlyingShotsman Thanks very much for all your comments which are all valid. With regards your comments: 1. Prior to doing the video I did check the voltage with the fluke at the output of the module, but as I was using 14 SWG wire of only short length about 10 inches I found that it made virtually no difference. You do have a point about the error though with regards compared to the fluke, but to be fair I thought for this type of module it was quite good. Since doing the video I have had a closer look at the resistor network they use to sense the input and output voltages for use by the 8 bit Micro - the resistors I said were 300 ohm were in fact 2k 1% (they used the EIA-96 coding of 30B and I miss read that to read 300). I also found if you adjust the value of the top part of the divider which is a 30K ohm you can trim the meter reading to read the same as the Fluke. I guess they did not do this due to the extra cost. 2. I agree with you about adding a heatsink to the bottom of the PCB. 3. Yes I agree it would be better if the display and LM2596 were held off until the last set voltage by the user was reached. The Micro chip remembers the last user setting in its internal memory. I have contacted the manufacturer/supplier and indicated to them they could make some improvements by simple modifying the code in the Micro. 4. Yes you are spot on about not using the croc-clip ground lead on the scope for best results. In my video I was just wanting to display the ripple and switch spikes to illustrate a point about Buck Converters. Later when I did used the earth spring on the scope probe as you suggested I found that it had most effect on the high frequency switching spikes which are in the Mhz region. The actual sawtooth ripple of the output capacitor charging and discharging was very much the same as what I showed on the video. Thanks for your kind remarks about my video which is most welcome. Again thanks for all you comments which I agree with entirely. Regards, Louis
Another excellent video. You reverse engineered an impressive schematic for it, with thorough explanation. I too received a DROK device for review, a very accurate little panel meter. I think DROK is really trying to get their devices in front of the hobbyist.
Corgitronics Thanks. Yes agree they produce some nice little modules for the hobbyist. I purchased my module on eBay and then found it was delivered by Amazon. So I looked in to it further and found the DROK website and the link they use to their Amazon Store.
Thank you for your channel! Could you help me with Ripple in input. I have 24AC and 12AC and I am using diode bridge with smoothing capacitor. For make input DC voltage for my LM2575. In datasheet I always see UNREGULATED DC INPUT. What must be Ripple peak to peak voltage befor DC-DC convertor? are there any rules or requirements for UNREGULATED DC INPUT? Thank you!
Hi, With regards an Unregulated DC Power Supply their output voltage will fluctuate with the input voltage from the AC supply and will also fluctuate depending on the load applied, in other words, the output voltage is not set to a fixed value. Furthermore, the output voltage is not fully smoothed. Therefore the output DC voltage of an unregulated supply will still fluctuation (ripple), the value of which changes depending on the load current. This should not be an issue for a DC-DC converter such as a LM2575. In the LM2575 datasheet there are details of the Input Capacitor Selection (Cin) for the DC-DC converter itself and this is separate to the unregulated DC Supply you are using, which states a 100uF aluminum electrolytic capacitor located near the input and ground pin will provide sufficient bypassing. This capacitor should have a low ESR (Equivalent Series Resistance) value. Its not the ripple that is important here, it is how low is the voltage allowed to sag in order not to violate the dropout specification for the regulator or in your case the LM2575. In short you should not be concerned to much about ripple from an Unregulated DC Power Supply. Simply choose a smoothing capacitor which is large enough to handle the maximum current you intend to draw from the DC-DC converter so as to avoid any voltage sag under load current on its output voltage. With an unregulated power supply powered by an AC supply (50Hz) and assuming full-wave rectification is used, the smoothing capacitor for your unregulated DC Power Supply must be able to supply current to the regulator/LM2575 for half an AC cycle (10 milli seconds if it is a 50Hz AC supply). So an example calculation could be: For a maximum 1A supply load current with a 50Hz AC supply, full-wave rectification, where you can tolerate say a 2V sag in the smoothing capacitor voltage without effecting the regulation of the LM2575. The calculation would be as follows: Value of smoothing capacitor (in Farads) = Maximum Load Current in Amps x (AC half cycle time in seconds/voltage sag allowed); this would equal in our example: Value of smoothing capacitor (in Farads) = 1 * (0.01/2) = 0.005Farad = 5000uF So in this example I would used the nearest standard value capacitor which would be a 4700uF. Hope this helps, Regards, Louis
Deian Cosas Thanks Deian. At 1.5 amp I found the temperature to be about 50°C. The module seems to handle up to 1.5 amp OK with out a heatsink. However, if your using the module for long periods at 1.5 amp or more I would recommend adding a small heatsink with thermal glue.
Robert Calk Jr. Thanks. I see your eagle eye spotted my Peak Atlas Analyzers in the background! The units are all very light so I have no problems hanging them by the leads. Makes them easy at hand on my workbench while been out of the way of getting damaged.
What a nice little circuit :-D, i like your diagram, makes it so easy to understand :-D. i did notice the control buffer op amp, a parallel capacitor across the gain resistor to limit the cut off frequency. I didnt expect the volt meter to be too accurate, but its good enough for ordinary use. The 431 shunt regulator ive seen in just about every switchmode psu made, placed on the output to drive an optocoupler feedback error signal to the pwm chip, simple :-). Very common in laptop 240 to 19v adaptors, easy to tweak as well. Thanks for a smashing review :-D
zx8401ztv Thanks for your comments. The capacitor in parallel with the gain resistor in the second op amp with prevent transients from causing excessive overshoot or oscillation of the output voltage of the DC-DC Converter. So you are correct in that it will limit the gain at high frequencies. The accuracy of the voltmeter in mainly dependent on the tolerance of the resistor potential divider, if you were to change the 30K resistor to say a 27K resistor in series with a 5K multi turn trim pot you could calibrate the meter and make it more accurate (you would also need to do this mode on both input and output divider networks). This is something I might do on another project I have in mind. The 431 programmable shunt regulator is a nice device and very cheap, I have used it many times in projects.
+Scullcom Hobby Electronics Is first part of LM358 correctly drawn on whiteboard? Is that should not be active low pass filter? How it is drawn it have gain same as open loop gain of LM358. To me it look like second-order Sallen-Key LPF but with 1k resistor and connection between it and capacitor drawn wrong.
Oh if only my white board diagrams looked so nice! ;-) Thanks for the overview Dave; DROK makes some other interesting stuff, a USB power meter and a USB 1A/2A "dummy load" that I've used here at my bench (also available on Amazon).
Nice review, but one question remains not clearly answered: Does it maintain the output voltage setting after the input voltage it turned off and on again?
By the way, it would be interesting if you did a video on filtering the output from these DC-DC converters to remove the high frequency spikes. Thanks.
Fake LM2596 chips have been reported to be unable to switch above 100 Khz frecuencies. Can you confirm that this one in particular is actually switching at 150 Khz?
The 150KHz oscillator is internal to the LM2596. The main advantage of a higher frequency switching oscillator is that you can use smaller value filter components. I do not know if the Drok uses a fake LM2596. All I can say is that it seemed to work as you would expect and in any event you could always add addition filter components on the output if the switching frequency was an issue.
The circuit I used was an overview so I did not show the power connections to the LM358. But I did indicate at the output of the 5V regulator that it was supplying power to the LM358 (indicated in RED).
ty. i looked at this video many times (my mind was on a psu build, base on arduino controlling 4x of xl2596) my bad didn't pay attention, and all your vids have description on very very details. last Q, what is the value of the 3x capacitor from lm358? again ty
I learned nothing from this video. You wasted 40 minutes of my time, performing tests and not really showing anything detailed and useful. Temperature is the most important variable in electronics, and yet, despite having a fluke probe, you rushed and kept switching the load off before the temp had a chance to stabilize. As a result, I never learned the point where this module can be operated without a heat sink. Same story when trying to determine the shut down or brown-out voltage. You jumped from 4 volts down to 3 volts. This is unacceptable characterization in anyone's book.
In practice this module can handle output currents up to about 1.5 Amp. Above that level you would need to add an external heatsink. For detailed information on junction temperature with varying loads and voltages check out the Texas LM2596 datasheet.
Yes, it was great to see the reverse engineering and the way you could build the schematic back. More of these skills and how to achieve this level of competence will be of great value. Thanks a lot!!
Wow.....what a fantastic and professional review, analysis and explication! Many thanks. I have a few of these in various guises but haven't got around to using them on a project yet. At least now you have got me more up to speed on them. Thanks again!
Glad you found it helpful.
Regards,
Louis
Fivefishdiy asked if there is short circuit protection. Although the LM2569 has thermal shutdown and current limiting protection I found in practice that shorting out the output tends to damage the LM2596 and it dies. I have been looking at ways of adding better true short circuit protection in a future project and plan to use the Linear Technology chip LTC1477 which has built-in short-circuit protection up to 2 amps.
+Scullcom Hobby Electronics I think it may well happen that cheap Chinese regulators use fake lm2596 chips that are not in par with branded chips. At least I saw fake power transistors on youtube and other fake ICs like lm7805. So I never buy ICs from China.
+Alexandre Kandalintsev
Same.
Someone I know spent thousands of dollars on hunks of plastic (fake ICs) from china ... they literally did nothing!
+Benjamin “Ozias” Esposti did you manage to get your money back? (hope so)
Alexandre Kandalintsev
No they didn't, because they bought from some random seller in china, after they bought the chips, the seller disappeared ...
Having had a closer inspection at the resistor values of the potential dividers at the input and output to measure the input and output voltage the resistors I suggested were 300 ohm are in fact 2Kohm 1% resistors. It seems that these resistors are marked as 30B (using the EIA-96 coding for surface mount resistors having a tolerance of 1%). All the other resistors on the module use standard 3 digit coding system.
Thank you so much for taking the time to draw the schematic and explaining it.
+Shaunak De You are welcome hope its helpful.
As usual, a very thorough review. Keep up the good work! It's very much appreciated.
Steve Rodgers Thanks Steve.
Great review, thanks for taking the time to draw and analyze the schematics.
One puzzling thing though: the 5V reg XC6203 seems to be rated at a max input voltage of 12V only, according to the datasheet.
LANE Boys RC Yes you are correct. I did wonder myself if the XC6203 would be damaged at higher input voltages but so far it seems to be O.K. If could be that there are different voltage rating types of the same IC and I will investigate this further.
Best review of this sort of device that I've seen. Especially your reverse engineering, very nice schematic and explanation of the unit's operation. Just a few comments. (1) You mentioned several times that it meets its specifications for output voltage, but in fact it never did. The closest I saw in the video was +110mV as measured by your Fluke DMM, and most of the time is was +150mV or more. Since your Fluke was sampling at the input to your DC load, it was actually even worse, if you account for the voltage drop in the wires from the Drok to the load. This should be measured at the Drok board output. (2) Adding a heatsink to the top of the LM2596 (as is commonly advised by the sellers of these devices) will actually do very little. The plastic case has very high thermal resistance and is not intended to dissipate heat. Surface mount power devices like this one use the metal tab on the bottom, soldered to the PCB, for that purpose. A heatsink stuck to the bottom of the PCB would be more effective. See page 25 of the Texas Instruments LM2596 data sheet. (3) The low voltage behavior of the device was pretty shocking. If the display is on, the output should match the display, and it didn't. The ST uC should be sampling the input voltage and holding the LM2596 enable low with the display off until the minimum dropout is met and the device can produce the displayed output. Only then should it switch on and show a value. (4) When measuring ripple on a switching converter, the high frequency components of the signal can make the measurements look a lot worse than they are, particularly if probing is done incorrectly. You really shouldn't use the croc-clip ground lead on the scope probe for anything above a few kHz. The probes come with a little spring that fits over the ground barrel if you remove the spring-loaded hook. The tip of that spring goes to ground and the probe tip to the node you're measuring. This greatly reduces inductive noise in your measurements. / Anyway, thanks for a very nice video. Cheers!
FlyingShotsman Thanks very much for all your comments which are all valid. With regards your comments:
1. Prior to doing the video I did check the voltage with the fluke at the output of the module, but as I was using 14 SWG wire of only short length about 10 inches I found that it made virtually no difference. You do have a point about the error though with regards compared to the fluke, but to be fair I thought for this type of module it was quite good. Since doing the video I have had a closer look at the resistor network they use to sense the input and output voltages for use by the 8 bit Micro - the resistors I said were 300 ohm were in fact 2k 1% (they used the EIA-96 coding of 30B and I miss read that to read 300). I also found if you adjust the value of the top part of the divider which is a 30K ohm you can trim the meter reading to read the same as the Fluke. I guess they did not do this due to the extra cost.
2. I agree with you about adding a heatsink to the bottom of the PCB.
3. Yes I agree it would be better if the display and LM2596 were held off until the last set voltage by the user was reached. The Micro chip remembers the last user setting in its internal memory. I have contacted the manufacturer/supplier and indicated to them they could make some improvements by simple modifying the code in the Micro.
4. Yes you are spot on about not using the croc-clip ground lead on the scope for best results. In my video I was just wanting to display the ripple and switch spikes to illustrate a point about Buck Converters. Later when I did used the earth spring on the scope probe as you suggested I found that it had most effect on the high frequency switching spikes which are in the Mhz region. The actual sawtooth ripple of the output capacitor charging and discharging was very much the same as what I showed on the video.
Thanks for your kind remarks about my video which is most welcome. Again thanks for all you comments which I agree with entirely.
Regards,
Louis
Another excellent video. You reverse engineered an impressive schematic for it, with thorough explanation.
I too received a DROK device for review, a very accurate little panel meter. I think DROK is really trying to get their devices in front of the hobbyist.
Corgitronics Thanks. Yes agree they produce some nice little modules for the hobbyist. I purchased my module on eBay and then found it was delivered by Amazon. So I looked in to it further and found the DROK website and the link they use to their Amazon Store.
@@scullcomhobbyelectronics1702 Modules from ebay == chinese factory seconds! That crooked input terminal block would drive me nuts to look at...
Thank you for your channel!
Could you help me with Ripple in input.
I have 24AC and 12AC and I am using diode bridge with smoothing capacitor.
For make input DC voltage for my LM2575.
In datasheet I always see UNREGULATED DC INPUT.
What must be Ripple peak to peak voltage befor DC-DC convertor?
are there any rules or requirements for UNREGULATED DC INPUT?
Thank you!
Hi,
With regards an Unregulated DC Power Supply their output voltage will fluctuate with the input voltage from the AC supply and will also fluctuate depending on the load applied, in other words, the output voltage is not set to a fixed value. Furthermore, the output voltage is not fully smoothed. Therefore the output DC voltage of an unregulated supply will still fluctuation (ripple), the value of which changes depending on the load current. This should not be an issue for a DC-DC converter such as a LM2575.
In the LM2575 datasheet there are details of the Input Capacitor Selection (Cin) for the DC-DC converter itself and this is separate to the unregulated DC Supply you are using, which states a 100uF aluminum electrolytic capacitor located near the input and ground pin will provide sufficient bypassing. This capacitor should have a low ESR (Equivalent Series Resistance) value.
Its not the ripple that is important here, it is how low is the voltage allowed to sag in order not to violate the dropout specification for the regulator or in your case the LM2575.
In short you should not be concerned to much about ripple from an Unregulated DC Power Supply. Simply choose a smoothing capacitor which is large enough to handle the maximum current you intend to draw from the DC-DC converter so as to avoid any voltage sag under load current on its output voltage.
With an unregulated power supply powered by an AC supply (50Hz) and assuming full-wave rectification is used, the smoothing capacitor for your unregulated DC Power Supply must be able to supply current to the regulator/LM2575 for half an AC cycle (10 milli seconds if it is a 50Hz AC supply). So an example calculation could be:
For a maximum 1A supply load current with a 50Hz AC supply, full-wave rectification, where you can tolerate say a 2V sag in the smoothing capacitor voltage without effecting the regulation of the LM2575. The calculation would be as follows:
Value of smoothing capacitor (in Farads) = Maximum Load Current in Amps x (AC half cycle time in seconds/voltage sag allowed); this would equal in our example:
Value of smoothing capacitor (in Farads) = 1 * (0.01/2) = 0.005Farad = 5000uF
So in this example I would used the nearest standard value capacitor which would be a 4700uF.
Hope this helps,
Regards,
Louis
Very good explanation of how the circuit works. I wonder how many degrees C would the module handle at around 1-1.5A.
Deian Cosas Thanks Deian. At 1.5 amp I found the temperature to be about 50°C. The module seems to handle up to 1.5 amp OK with out a heatsink. However, if your using the module for long periods at 1.5 amp or more I would recommend adding a small heatsink with thermal glue.
Thanks for the review. You aren't afraid of hurting the leads hanging your Peak Atlas Analyzers like that?
Robert Calk Jr. Thanks. I see your eagle eye spotted my Peak Atlas Analyzers in the background! The units are all very light so I have no problems hanging them by the leads. Makes them easy at hand on my workbench while been out of the way of getting damaged.
Scullcom Hobby Electronics
I bought the hard cases for mine so they stay clean and safe.
Great video as always, and impressive schematic drawing!! Reminds me of the neat hand-drawn schematics in old service manuals. :-)
Darian Cabot Thanks Darian.
What a nice little circuit :-D, i like your diagram, makes it so easy to understand :-D.
i did notice the control buffer op amp, a parallel capacitor across the gain resistor to limit the cut off frequency.
I didnt expect the volt meter to be too accurate, but its good enough for ordinary use.
The 431 shunt regulator ive seen in just about every switchmode psu made, placed on the output to drive an optocoupler feedback error signal to the pwm chip, simple :-).
Very common in laptop 240 to 19v adaptors, easy to tweak as well.
Thanks for a smashing review :-D
zx8401ztv Thanks for your comments. The capacitor in parallel with the gain resistor in the second op amp with prevent transients from causing excessive overshoot or oscillation of the output voltage of the DC-DC Converter. So you are correct in that it will limit the gain at high frequencies. The accuracy of the voltmeter in mainly dependent on the tolerance of the resistor potential divider, if you were to change the 30K resistor to say a 27K resistor in series with a 5K multi turn trim pot you could calibrate the meter and make it more accurate (you would also need to do this mode on both input and output divider networks). This is something I might do on another project I have in mind. The 431 programmable shunt regulator is a nice device and very cheap, I have used it many times in projects.
+Scullcom Hobby Electronics Is first part of LM358 correctly drawn on whiteboard? Is that should not be active low pass filter? How it is drawn it have gain same as open loop gain of LM358.
To me it look like second-order Sallen-Key LPF but with 1k resistor and connection between it and capacitor drawn wrong.
Oh if only my white board diagrams looked so nice! ;-)
Thanks for the overview Dave; DROK makes some other interesting stuff, a USB power meter and a USB 1A/2A "dummy load" that I've used here at my bench (also available on Amazon).
Dino Papas Thanks Dino. I will have a look at their dummy load.
Louis
Scullcom Hobby Electronics
Oooops, sorry Louis, brain was definitely not engaged last evening and I mistakenly addressed you as "Dave". ;-)
Nice review, but one question remains not clearly answered: Does it maintain the output voltage setting after the input voltage it turned off and on again?
Chip Guy Vids Yes it remembers the last output voltage set by the user. It stores the information in the Micros eeprom.
Nice little unit.
Thanks for the review...keep them coming .
Dave.
Dave Bassett Thanks Dave.
Excellent video, very thorough. I'm looking forward to seeing more of your videos.
By the way, it would be interesting if you did a video on filtering the output from these DC-DC converters to remove the high frequency spikes. Thanks.
Extremely useful and informative video. Thank you very much.
Thank you professor, great material All the best.
Thanks diX.
How about a video on tracing out a PC board and creating a schematic like in this video. Just a suggestion, keep the content coming.
+rpm750 Thanks for your comments. I may consider your suggestion for a later video.
Awesome explanation! Thank you for sharing.
+saturn5tony Thank you.
Thank you for a great review my friend. Have a great week.
Anybody home? I need a hand finishing something. Trying to build something yet no channels are still live and/or ever respond.
Is there short circuit protection?
Thanks for the review and excellent explanation :-)
Mark Garth Thanks Mark.
Fake LM2596 chips have been reported to be unable to switch above 100 Khz frecuencies. Can you confirm that this one in particular is actually switching at 150 Khz?
The 150KHz oscillator is internal to the LM2596. The main advantage of a higher frequency switching oscillator is that you can use smaller value filter components. I do not know if the Drok uses a fake LM2596. All I can say is that it seemed to work as you would expect and in any event you could always add addition filter components on the output if the switching frequency was an issue.
Thank you.
Great video thanks. Good work on the reverse engineering.
+netdudeuk Thank you
LM358 is power up from???? input? output?
The circuit I used was an overview so I did not show the power connections to the LM358. But I did indicate at the output of the 5V regulator that it was supplying power to the LM358 (indicated in RED).
ty. i looked at this video many times (my mind was on a psu build, base on arduino controlling 4x of xl2596) my bad didn't pay attention, and all your vids have description on very very details. last Q, what is the value of the 3x capacitor from lm358? again ty
I learned nothing from this video. You wasted 40 minutes of my time, performing tests and not really showing anything detailed and useful.
Temperature is the most important variable in electronics, and yet, despite having a fluke probe, you rushed and kept switching the load off before the temp had a chance to stabilize. As a result, I never learned the point where this module can be operated without a heat sink. Same story when trying to determine the shut down or brown-out voltage. You jumped from 4 volts down to 3 volts.
This is unacceptable characterization in anyone's book.
In practice this module can handle output currents up to about 1.5 Amp. Above that level you would need to add an external heatsink. For detailed information on junction temperature with varying loads and voltages check out the Texas LM2596 datasheet.