I have a question here about a major blunder of mine. Back in May or June I started building the Cadet modules and today, while I was working my whole system just stopped working.
The output just became static. I started pulling out my modules and I noticed that on my Cadet 2 I never placed the 7805 voltage regulator. So, I quickly went and soldered one in and now…nothing. This is probably a dumb obvious question but, is it worth trying to trouble shoot it? Or do I just have to go ahead and replace the RGB Encoder? It worked well for months, the only issue I had was with getting the ramps module to work. I can’t believe I forget to put the voltage regulator in. Thanks.
That seems strange it was operating without the 7805 at all since that’s what powers three of the ICs in the circuit.
Can you post a hi res photo of the front and back of your board?
Here are photos, keep in mind though it was working properly (to the best of my ability to tell anyway) for a few months. I added the voltage regulator in, after I realized it wasn’t there.
Hmm, I wonder how the ICs that run on 5V, especially the AD724, were being powered to encode the composite signal.
Do you have a multimeter to check if you are getting +12V/-12V/+5V on the power pins of each of the chips? This is probably the best place to start debugging. If you look up the datasheets for each IC, it will tell you which pin to check but just be careful not to short pins when probing.
A good number of your solder joints could use reflowing, too. I’m not certain if some of those are cold joints because the surface was disturbed with how you trimmed the excess component leads. It is important to keep your component leads sticking out of the board short but you can potentially disturb a solder joint if you cut into them too much. You should aim for a smooth, volcano/Hershey’s Kiss shape on every joint.
I checked the IC’s and the AD724 is coming at 5V as well as the 74HC14 and CD4053. The LM6172’s are reading at 22V which seems high.
Okay, that’s good you’re getting 5V on those chips.
You should be seeing something like ~11-12V on each side of the LM6172s supply pins so that kind of sounds like you’re possibly measuring the potential between the V- and V+ pins on the LM6172s so you’re seeing 22V peak to peak. Do you get something between +11-12V when you reference the V+ pins to ground and -11-12V when you reference the V- pins to ground?
Are any of the chips warm to the touch?
No, I was measuring the V+ and V- pins. The IC’s weren’t hot. Any other ideas what I should check?
If you’re getting +22V and -22V on V+ and V- referenced to ground, there’s something really wrong there. LM6172s can only take up to 36V peak to peak so if it’s somehow getting 44V peak to peak, they’re likely fried.
Can you confirm the V+ / V- power supply voltages referenced to ground (the 7805 tab is a good ground reference) by probing the legs of the LM6172s rather than the pads they’re soldered to, if you weren’t already doing that in your measurement? If you can measure what’s on the Eurorack power header as well to confirm you’re getting close to the expected +12/-12V, that would help narrow it down too.
Nothing is jumping out to me as wrong component placement but I think debugging attempts would benefit from methodically and slowly reflowing almost of the joints on the board. There’s definitely a few pins that look like they don’t have enough solder, especially those that look flat/dull/convex instead of a neatly sloping shiny concave cone. This doesn’t mean they aren’t actually soldered but it’s going to be really difficult to tell without probing every single node with an oscilloscope if they don’t look like that. You might need to turn up the heat on your soldering iron or use a little bit more solder.
I’m still perplexed how it was working before with no 5V supply but I’m sure we can puzzle it out if we probe enough. You don’t happen to have an oscilloscope, do you? We can verify output on a couple key components with that but a DMM won’t tell you too much for video rate signals. I’d recommend carefully reflowing the board before we go down that route.
Generally, if you’re certain it worked without the 5V reg installed, more or less the only possibility is for it to have been powered through the I/O pins of one or more of the 5V parts.
Almost all IC’s (Some exceptions for analog parts, you know these by their extra aggressive esd warning stickers) have ESD protection on the non-power pins. THis takes the form of a diode going down to VCC/VDD (i.e. the upper supply, so the diode is normally blocking) and up to VEE/VSS (or ground if no lower supply exists).
This usually serves to absorb small voltage spikes into the power supplies rather than frying the transistors on the pin.
More relevant to this case, this also allows you to power a pin (poorly, but enough to work) through the non-power pins.
Depending on exactly what powered what through what, that might have worked until something that was not meant to provide power gave in (or something else changed.)
On C2, the only ICs on the 5V bus are as far as i can see [U4,U5,U7].
U4 seems to be connected only to the other 5V ic’s and to the output (and thus cannot be our culprit)
U5 seems to be connected to the other 5V’s and to the RGB signals (unlikely to be powered through these)
U7 is connected to the sync bus.
If there was no shorts in addition to the missing part (and the apparent normal functionality of the module supports this) the only obvious part to me is the sync bus.
either way, i’d swap out U7 and test.
If that does not work, i’d try to swap U5 (being the only other easily swappable part) and if possible the buffer part on the sourcing side of the sync bus (U7 on C1, i think).
+/-22V sounds like V+ to V- after the polarity protection diodes
I was incorrectly testing the LM6172’s which was why I was getting the 22 V readings. When I used a proper ground I was getting a reading of 10 or 11 V.
I am absolutely certain that the RGB Encoder was working without the 7805. I went on to build another 5 Cadet modules, and 3 Castles and they were all successful, with exception of Ramps where I could get the the horizontal ramp to work but not the vertical (which may be related to this).
I had just completed the Cadet 3 and was testing that when this happened. While I had a patch running, I believe I had the Cadet VCO going to the blue on the encoder and it slowly went from the pattern it was making to a full blue screen and then static. Which felt like a chip failing. I started taking out all my LZX modules and that is when I saw the missing 7805 on my RGB encoder.
I went ahead and replaced the 74HC14 and CD4053. Right now, I am only running my Cadet 1 and 2 and when I plug in the my Cadet 2, I get static even with the two new IC’s.
I checked the out put of the RGB Encoder on my scope and this is what I am seeing. Just got my scope so I am still learning how to use it, so I hope I measured this correctly.
Here’s what I am getting on the oscilloscope:
Here’s better image of the waveform:
Are you sure you have proper grounding for your probe? the 10Vpp AC signal certainly does not look right for video, but does not really look right for any plausible error mechanism either. it does however look quite right for a floating probe/ground hum.
If you’ve swapped the analog switch and inverters, my gut feeling is that either it’s the AD724 (i.e. extremely hard to fix), or the fault was unrelated to the 5V issue.
- If you connect it to a PAL-compatible monitor and switch both the sync generator and the C2 to pal, do you get the same error, a different error, or no error?
- If you put some constant signal into the blue input and scope one of the pins pins of C14 you should see a low-voltage square wave.
- What does the signals on the s-vid connector (J7) look like?
Thanks to both of you for guiding me through this. I think I am going to try to build another Cadet 2, but I would love to determine what the problem is on this guy for my own education.
I do have a square wave coming from C14:
My gut feeling is the AD724 is shot too.
The output of J7 looks like a sine wave:
If you don’t have a pal-compatible monitor it would still be useful to see if the signal still looks like a 60Hz wave on the output with both switches set to PAL, but yes, it may well be fried.
I ordered and built another RGB Encoder and all is working again. I am convinced AD724 got fried. I didn’t skip the 7805 this time.
Sorry the AD724 probably died but great news you’ve got a working Encoder! If you want to attempt to remove the chip and replace it, you can get free AD724 samples (along with lots of other interesting chips!) from Analog Devices’ website.
PM me if SMT rework isn’t your thing.