EEZ Bench Box 3 (BB3)

[prasimix]

<snip>

<snip>

Remove 16-pin IDC cable and check if standby LED is active when turn power on. Few other backers reports that cable failure. Also you should hear fan is spinning for a few seconds after the power is turned on.

Ah, yes, that gives fanspin and a green standby LED. Thanks!

I had removed the AUX-PS board and was checking it out and I could not find anything that was wrong (without connecting AC power that is), even the power button checked out ok.

I will bring out suitable probes and check that cable in detail...

Ah yes, all pins are connected to all other pins, probably not what was intended 

The attached image is a bit fuzzy, but it can be seen that the cables are not in their "grove" but rather right in between so the "knifes" or whatever one wants to call them cuts through the cables on either side of it so everything gets connected together. Quite an achievement that...

If this was a TME part it probably is not the best place to get a replacement?

Send me your address, and I'll send you a new cable. TME has one but it is a little bit longer 15 cm: https://www.tme.eu/hr/en/details/fc16150-s/ribbon-cables-with-idc-connectors/amphenol/

[prasimix]

I would rather suspect the DCP405 modules themselves. Can you check if the same problem exists when the current range is set to High (see Section 11.5.1)?

No, can't get the issue when setting it to high range. Going down in the 10s of uA range makes the CC LED pulse slowly. If it wasn't a fault it would actually be a nice "heartbeat effect":
The limit values were only reduced to get a better visibility on the graph. They don't seem to affect the behavior.
Maybe this slowed down version helps diagnosing the behavior. If I can provide any other information, please tell me!

Kind regards,
Max

Ok, so there is a good chance that something is not soldered properly (or something dropped off) in current control loop section. Perhaps R62, IC21, etc. Please let me know if you can spot anything suspicious, otherwise we could organize repair. Sorry for that.

[MaxZ]

Mh haven‘t spotted any issues so far... I did gently push a few suspicious components with a screwdriver but nothing moved.
I took pictures of the area; I dont have a microscope for this so my phone had to do the job. Far from perfect but maybe not completely useless. This is btw. the left/first module (the one I already had to solder on).

Thumbs up for the shematic as PDF which I could search through for the component. Would be great to have a PDF of the PCB as well - or did I simply not see it?

Edit: Those faston connectors of the ground wires are probably the tightest ones I‘ve ever had. I can‘t unplug the one of the second DCP405 module with my fingers. I don‘t have slippery fingers at all; if I pull harder I‘ll just hurt myself. So yeah; will have to use pliers... Just to get this right, I totally agree that they should by no means sit or become loose. However I don‘t think this is a connector thats made to be used with pliers and I also worry about putting that much mechanical stress on its solder joints...


Kind regards,
Max

[prasimix]

Thumbs up for the shematic as PDF which I could search through for the component. Would be great to have a PDF of the PCB as well - or did I simply not see it?

PCB top and bottom side generated by OSHPark can be found in Images folder.

Thanks for pictures, that section look good, question is does IC21 is working properly. Another section of interest is around IC10. It is possible that some of the components do not have a good value. That “heartbeat” looks like some oscillation when it’s in the low range and it’s possible that something is wrong with the compensation on those modules.

[Kleinstein]

For the voltage not reaching the set value in the low current ranges, it could be some kind of oscillation. With some 5 mA this could be faster, so it would not be visible as a heart beat, but still cause the on average lower voltage.
It may be worth checking the current reading in the low range - it could be a problem with the shunt switching so that the low shunt has not connection.

The down-programmer could be an error source that could be turned off for a test.

With 2 such broken boards the test procedure should ideally include a test that could detect such a failure. With better solder quality such errors are less likely, but one never knows.
 

[prasimix]

Good point @Kleinstein, one of the possibility is that driving of Q10 is faulty, and when in low range mode current flows via D14 on which voltage drop is higher then on Q10 and CC compensation is not effective.

[prasimix]

With 2 such broken boards the test procedure should ideally include a test that could detect such a failure. With better solder quality such errors are less likely, but one never knows.

We'll definitely need to include more testing for low range mode.

[goran.mahovlic]

Thumbs up for the shematic as PDF which I could search through for the component. Would be great to have a PDF of the PCB as well - or did I simply not see it?

Hi Max you can try this gerber view

https://tracespace.io/view/?boardUrl=https://kitspace.org/boards/github.com/eez-open/modular-psu/dcp405/dcp405-f688500-gerbers.zip

This is also amazing software, but someone would need to add boards and calibrate them...

https://www.inspectar.com/

[IanJ]

Hi,

I got intrigued by the bad heartbeat issue and powered up my BB3 to have a test, and made a short video of my experience.

https://youtu.be/649JsJ7pv3g

Ian.

[MaxZ]

Thanks for the inputs!


Unfortunately the pictures and the gerber viewer don't serve the purpose I had originally in mind. This being a completely unknown PCB to me, I wanted to search for the component name to find it quickly. Not very important though.

Looking at IanJs video I note a few things:
 * he has a similar issue, although at lower currents.
 * Recovery goes very fast until it almost reaches the set voltage, then it takes considerably more time. Is that normal...? At least that's something I can't see on mine. But I can't go that low in current without hitting the heartbeat issue either...Edit: Sure, same behavior on mine, too. Still: is that normal?
 * We're now at at least 3 DCP405 with pretty much the same issue @IanJ: how about your other DCP405?

Progress on my side!
As far as I can see I have no missing component on the entire board when comparing it to the PCB images on GitHub. For quickly checking component orientations it would be nice to have a high resolution picture without heatsink.
Disabling the down-programmer "solves" the issue! I attached two screenshots that show the output voltage after enabling the output, one with down-programmer enabled, one with it disabled.

Kleinstein suggested to change the current reading to the lower range, too. I don't know how to do that? I have set the DCP405 explicitly to the 50mA current range but I'm still getting only 0.5mA resolution on the read out value - even if my set value is 10uA.


Kind regards,
Max

[Kleinstein]

So it looks like the problem with CC to CV cross over at low currents is not so unique, just the current where it happens can be different.

In the circuit there is the part with Q27 that could be part of the problem. It seems to turn down the set voltage a little (some 10%) when in CC mode, with the idea to limit voltage overshoot when recovering. With a very low set current this could cause trouble: when no longer in CC mode Q27 turns off and raises the voltage. With a very low current setting, the voltage regulator part may call for more than the limiting current to charge the output capacitor. This would cause to get back into the CC mode. This could cause the slow oscillation / hart beat.

I see another problem with this circuit part: with a current limit just a little lower than needed to get the full voltage with a resistive load (e.g. 10.1 V and 10 mA setting with 1000 Ohms load resistance) the CC mode would engage and set the voltage limit down below the correct 10 V value.

For the low current range one could check the current readings. E.g. get some 10 mA and check with an external DMM - if the range switching is bad the current could be much lower (e.g. nA to  µA range) - though probably more current from leakage. So if the shunt switching does not work, it would probably be way off.

[MaxZ]

Update: Disabling the down programmer removed the oscillation but unfortunately did not completely solve the issue... Enabling the output with 10V/20uA makes the output rise linearly (CC LED only) up to 9.76V, where it'll stop with both CC and CV LEDs on. Not sure but this seems to match Kleinsteins explanation(?)

Kind regards
Max

[prasimix]

@Kleinstein diagnosis is accurate, the issue is related to Q27 which was subsequently added (hence the high designation number) to prevent overshoot while still maintaining a high rising time when the voltage increases. This in combination with the down-programmer generates a problem at low set currents. We can solve this part in software by turning off the down-programmer at very low currents (where it doesn't make much sense).

EDIT: question is what current level to set as a threshold for deactivating down-programmer.

EDIT2: This issue exist only in low current range.

[MaxZ]

As I said just before: this would remove the oscillation, but does not solve the issue with the output voltage not recovering completely...

Update: Disabling the down programmer removed the oscillation but unfortunately did not completely solve the issue... Enabling the output with 10V/20uA makes the output rise linearly (CC LED only) up to 9.76V, where it'll stop with both CC and CV LEDs on. Not sure but this seems to match Kleinsteins explanation(?)

Kind regards
Max

[Kleinstein]

How does the voltage output look with a scope - chances are in the non working CC/CV cross over case there would be some kind of oscillation, that may be a little to fast to be seen with the integrated voltage and current readings.

For a test if Q27 is part of the problem one could remove R118 or solder a bridge over Q27 (would make the voltage setting some 10% lower). As a side effect there may be more voltage overshoot - so it is not a final solution, just for testing. It would than need a different solution (e.g. larger value for C47, maybe C44+R51 instead of C46 or some limit to the output of IC9A) to limit voltage overshoot.

Because of the other expected issue (problem if the current limit needs the voltage to drop just a little), we may need an alternative to Q27 anyway.

[elekorsi]

Well after reading this i had to check mine and i can confirm that it acting the same on both modules. It doesnt recover if current is set below 7mA. I tried with the 100ohm resistor on output and it is the same. Disabling the down programmer eliminates this.

What i dont understand is, what hapened to my DMM. It was connected in series and i it blew the fuse!! When this happened, the voltage was set to 24V and current to 5,5mA, with down programmer disabled.
It is a Fluke 179 with the 440mA fuse (genuine one, as the meter was never opened before - i am the only one who uses it)...

[MaxZ]

@Kleinstein: No oscillations as far as I can tell. However, [edit]during ramp up and[/edit] when it is stuck in this CC/CV crossover, I do pick up more 50Hz noise (20mVpp, while open circuit [positive cable disconnected from BB3] noise is 40mVpp). Once it‘s in CV mode the noise is gone. A little bit strange but nothing I worry about; its not the job of the power supply to get rid of noise from unshielded lab cables.

@elekorsi: since it‘s not an assembly fault probably all DC405 modules are to some extend affected by this issue.


Kind regards,
Max

[prasimix]

Well after reading this i had to check mine and i can confirm that it acting the same on both modules. It doesnt recover if current is set below 7mA. I tried with the 100ohm resistor on output and it is the same. Disabling the down programmer eliminates this.

What i dont understand is, what hapened to my DMM. It was connected in series and i it blew the fuse!! When this happened, the voltage was set to 24V and current to 5,5mA, with down programmer disabled.
It is a Fluke 179 with the 440mA fuse (genuine one, as the meter was never opened before - i am the only one who uses it)...

So far I managed to blow 440 mA fuse only during the calibration when forgot to switch DMM input terminal when switch from low to high current range when second point is about 4.5 A

[MaxZ]

So far I managed to blow 440 mA fuse only during the calibration when forgot to switch DMM input terminal when switch from low to high current range when second point is about 4.5 A

Jep... I know what you mean... It might be a good idea to show a warning, or even simpler and inherently safe: swap the order; starting with 4.5A first, and then going down to 45mA (can't remember the exact value).  It is a tiny bit counter intuitive since we want to start at the left and go up step by step but if this comes at the expense of half of the user blowing their multimeter fuses...

Kind regards,
Max

[elekorsi]

I understand this, but to me it happened when the current was set to 5,5mA, not during calibration or any other misstake with the high current setting.

[prasimix]

I understand this, but to me it happened when the current was set to 5,5mA, not during calibration or any other misstake with the high current setting.

That could happen only if DCP405 has serious fault in CC control loop. If that is a case you can check that with same setup and DMM set to 10 A. It should notice 400+ mA current.

[IanJ]

I understand this, but to me it happened when the current was set to 5,5mA, not during calibration or any other misstake with the high current setting.

That could happen only if DCP405 has serious fault in CC control loop. If that is a case you can check that with same setup and DMM set to 10 A. It should notice 400+ mA current.

For info,

With a couple of small test cables and a 500mA fuse, and setting the BB3 to 40V/50mA I can blow the fuse if I touch the cable off and on a few times.
With a 1A fuse I can see the fuse wire within the glass move and return to centre when I attach the leads.
Touch the leads togeher without a fuse and you'll get lots of sparks.

CC mode reaction time, cable inductance, back EMF......not entirely sure but all my bench PSU's do this. There is a lot of energy to dissipate.



Ian.

[darmar]

Hi everyone,
...
- The cable for the fan is about 3cm too short? I can't put it the way as shown in the assembly manual. If I put it from the connector down between the DC/DC converters and the heatsink, there's not enough length to put it against the backplate; it goes straight to the fan. (think of a triangle. hypotenuse is shorter than the other two sides together). Is this meant to be like this? Again, doesn't cause any issues I think but didn't seem right to me...
...
Max

Figures 1 and 32 (https://envox.hr/eez/eez-bench-box-3/eez-bb3-building-instructions/bb3-kit-assembly-instructions.html) show the fan rotated 90 degrees clockwise.
The fan leads are on the left side, not the bottom side.

[prasimix]

CC mode reaction time, cable inductance, back EMF......not entirely sure but all my bench PSU's do this. There is a lot of energy to dissipate.

Not really understand this: you mean DCP405 behave differently or similar to your other PSUs?
We have a 33 uF elco on output that needs to be discharged and it is generally better to use OE ON/OFF control than direct load connection.

[IanJ]

CC mode reaction time, cable inductance, back EMF......not entirely sure but all my bench PSU's do this. There is a lot of energy to dissipate.

Not really understand this: you mean DCP405 behave differently or similar to your other PSUs?
We have a 33 uF elco on output that needs to be discharged and it is generally better to use OE ON/OFF control than direct load connection.

Yes, the DCP405 is the same as my other bench PSU's........so I don't really see a problem and it's probably why fuses such as the 440mA DMM fuse mentioned earlier can get blown despite setting CC to well below the fuse rating.
In my post/photo I just tried to demonstrate that.
33uF I guess would do it for sure....

Ian.