I’ve built the voltage ref plus a single channel of the Cadet II circuit on veroboard (U1, U2, U9) and I’m seeing my U2 getting extremely hot (after about 45 seconds it starts smoking) when I power it up. Has anyone DIY’d this circuit (based off the schematics, not the LZX PCB) and seen similar issues?
I’ve checked my layout/wiring several times (especially the supply pins), visually inspected AND checked for shorts with a multimeter, checked every net for connectivity and metered passive component connections between pins (with the chip out of circuit). I can’t find anything wrong. I tried two different LM6172s (purchased from Mouser) - same result both times.
I know it’s most likely still an issue with my layout, but I’m asking just in case someone has been down this road and had this problem before me 
Are you using all the same resistor values as the schematic and 1n5711 schottky diodes?
What are you using to generate your voltage reference? Is it from a buffered source?
The schematic is correct; it matches the PCB layout.
I am using 1N5711, LM6172 and TL431, all purchased from a reliable source (Mouser). I am using the voltage regulator circuit from the cadet II schematic (U11, U9). In my build I have shared the one LM6172 for both U9.2 and U1.1, in case that matters.
For the resistors, almost all are as specified in the schematic, excepting three that I substituted for values I had on-hand:
In the voltage reference circuit, R38 4.99K I used 5.7K and R40 3.48K I used 3.6K. By my calculations, 3.48/4.99 = 0.69 and 3.6/5.7 = 0.63 which is about 10% off, so my voltage reference won’t be quite right, but I wouldn’t think that would be the cause of my LM6172 overheat issues.
Between the U1.1 output and the U2.1 input, R13 1.4K I used 1.5K. A higher value resistor should attenuate the input signal slightly, but again I can’t see how this might cause the overheat issue.
Do you think the substituted resistor values are at fault? I do have the correct value resistors on order.
I’m using a DIY FCUK power supply, and the rails look clean and I have used it to power other circuits just fine.
Thanks again for your help.
Did you place 100nF bypass caps to GND right at the +V and -V pins of your opamps? The caps are in the schematic, but maybe easy to miss.
You’re right - I missed them. I will remedy this tonight (and try a new LM6172).
Hopefully that’ll do! Bypassing becomes more critical when you’re working with these wideband parts.
Cool, thank you. I’ve only done audio frequency electronics up until recently. Things do seem to be a lot more exacting with video!
1 Like
Thanks Lars, I can confirm that installing 100nF power conditioning caps to ground on all LM6172 supply pins in solved the overheating LM6172 issue. You are a legend, thank you!
Without them I guess there must have been undamped high frequency oscillation taking place, beyond the chip’s maximum operating specifications, causing it to heat up.
1 Like
That sounds like the correct diagnosis to me. Good work!