Cadet/Castle components sources and substitutions

As Thonk is not currently shipping to the US because of the pandemic, does anyone have another source for those 9mm B10k pots?

Pots are generally manufactured with specific options, and Thonk would have custom-ordered a large quantity of them to get them built specifically that way. They’re a great choice, but unfortunately not available anywhere else unless you order 20,000-odd from the manufacturer :slight_smile:

However, I have successfully substituted these sorts of pots:

You have to carefully bend the legs 90 degrees out from the body, and then very carefully trim back one side of the thick part of the legs with a dremel in order to fit the PCB holes. It’s easy to accidentally cut the leg off completely, so be careful (maybe buy a couple of spares). It looks a bit hacky, but it works just fine.

I’m sure there are other places you could buy similar pots locally. Check the shaft to make sure it matches your pots, and make sure the body is the right size (i.e. distance from pins to the front panel) by reading the datasheet. Always read the datasheet :wink: The ones I used were actually spline shaft, since that matched the knobs I had.

Explodingshed sells them. (based in Germany)

you can also buy 9mm vert pots on alibaba, ebay etc.

I looked at the ExplodingShed web site (thanks for the pointer Martijn) but couldn’t see ones that were quite the same as the Thonk ones? They did have a suggestion for modifying other pots, similar to the way I mentioned.

On ebay I did find this one which is pretty similar (but rather expensive):

I found these (not tested yet)

and cheap right angled

That last one looks very good - with a bulk purchase and combined shipping it would be a great price!

wait, this one is the one I meant:
I’ve ordered this one a few times already, in packs of 100

1 Like

Can anyone specify the bandwidth requirement that makes the LM6172 the only possible op amp? I understand that slow op amps within the video path will blur or muddy the picture, correct?
Mouser lists the LM6172 as having a GBP of 70MHz. The next opamp down in speed and cost is the NJM2068D with a GBP of 27MHz. This sounds like a big drop in bandwidth but I at the same time, 70MHz sounds like massive overkill when a colorburst signal is only 3.58- or 4.43MHz.

Correct me if I am looking at the wrong spec.

1 Like

the slew rate is also important


I haven’t used that chip before but the slew rates on the datasheets for both chips look significantly different (6V/µs vs. 3000V/µs on the LM6172).

You could pretty easily compare them if you just breadboard a buffer circuit like this (remember 100nf bypass caps on V+/V-) and feed it signals with hard edges like a keyed shape or hi frequency square wave oscillators. It should be readily apparent on the output if the NJM2068D isn’t slewing quick enough.

The 1k resistors on the op amps are suggested in the LM6172’s datasheet for buffers. Not certain if that’s required for the NJM2068D or if you can get away with a wire link like in a TL07x buffer circuit. Let us know what you find out if you try this!


It’s not the only possible opamp to use for video stuff. There are many fast opamps that will work, but it’s the one that LZX chose for their DIY designs. You may be able to use other opamps but they probably wouldn’t be a drop-in replacement into the existing Cadet boards.

I can’t speak for LZX, but if I had to guess why they chose it for the Cadet stuff, it’s probably because:

  • same pinout as the common TL07x/08x series
  • available in through-hole (DIP) packages
  • available from major electronics distributors
  • cheap(-ish) for a fast opamp meeting the other specs

Probably some other reasons too… maybe involving offsets or something else I can’t quite remember? The pinouts and the DIP packages especially mean that they’re easy for people to experiment on their own with and design new things, which I believe is something LZX wanted the Cadets designs to help with. And I know we don’t really think of them as cheap, but when you compare them to other video-rate opamps, they’re towards the lower (or lower-middle) end of the price spectrum (or at least they were when I looked a couple of years ago).

And the reason for the LM6172’s in the Castle line are probably all the same reasons as for the Cadet line, plus the fact that they were already used in the Cadet line (since Castle came later).


Here are the standards:

Slew rate is what you should care about when evaluating an op-amp. You want to look at the time it takes to go from 0V (black) to 1V (white). That time should be less than or close to 1x SD pixel (about 70 ns.)

LM6172 is getting more and more expensive since TI bought out National. But there aren’t any real alternatives with the same specs, price, available in DIP, and that can run on +/-12V power rails. We’re moving to SMT for our next rendition of DIY designs, and using the same opamps we use for production (ADA4851-x series, running on +/-5V power rails.)


Thanks for the replies, everyone.
The 70ns spec is essentially what I was after. I was able to find cheaper op amps with comparable slew rates but like Lars pointed out, the packages are different and most notably the max power supply ratings are lower.

Also much appreciated that no one has yet to speak down to me. What a refreshing community you have here.

@rempesm, I did toss a few of the 6v/us op amps into my cart along with all of the 6172’s that I need. At 54 cents each, I am curious just ‘how’ bad they can possibly look.


I notice that Rochester Electronics has a stock of the obsolete, not lead-free version of the LM6172IN for less than half the cost of the NOPB version.

Though the minimum qty is 120. (Another group buy?)


Yeah, you can get neat unintended effects with narrower bandwidth parts. Often it appears to be horizontal smearing or other filtering effects but you might get surprised. I usually stray away from IC sockets for video modules but you could try that on one of your builds to make comparisons easier. Breadboarding that buffer circuit and passing thru signals that sharply change from B<->W is probably the more direct way to roughly evaluate the difference. Breadboards will introduce some stray capacitance just like IC sockets.


Do you need a chunk of them atm? The difference for the 500 and 1000+ tier seems kind of negligible. I’d be down to split a 120 pcs order with someone in the US but don’t have bandwidth to arrange a huge group buy sendout.

1 Like

Sean, you’re a legend! Thank you so much for this!
Given I’m in Australia there’s not really any point in me sharing a group buy with you guys so I’ve just done my own order.
Now I just have to build up 120 x LM6172’s worth of video modules :slight_smile:


Now I just have to build up 120 x LM6172’s worth of video modules

OH no, what a disaster!

Good find! Put me down for 30-50. Maybe more. How many did you need? If we can get a group buy set up this week, that would be fantastic.

1 Like

Sorry, probably worded that wrong… I should have said, “Perhaps those that need some could arrange another group buy?”

Myself, I’m actually still working my way through my stock from the last group purchase.

But, yeah, I’d imagine getting a few people together to reach the 120 minimum shouldn’t be tough.