Yes, the board has worked out well. However, there is one mistake on the power supply regulator FET: the gate and drain pins are reversed. The FQP19N20 has the pins as G, D, S, whereas the board has D, G, S. Not a huge problem as you can just bend the D and G pins and make it fit. I am using the board like that right now and it works just fine. There might also be some FETs with the pins configured as D, G, S, though I haven't looked for options on Mouser yet. I will for sure fix this in the next iteration and I thought I had confirmed it in KiCAD, but clearly I missed it. You could also use your own PS and wire in the power right before the CCS but in my testing I didn't find any audible differences between the on-board supply and a fully tricked out Salas Shunt Reg.
Other than that, the PS and regulator works fine. I would also recommend not using the jumpers that shunt the zener references. I added this feature mainly for me to play around with the supply voltage, but it is not needed for a build. You can just put the zener stack in and don't wire any jumpers. The zener stack should total to about 100-130V. I'll add this to the notes on the schematic.
What's left to do:
1. Get PCBs out to those who are interested. I received six in the original order and I'd like to keep two. So, that leaves four extra. I'm going to ship these for free to whoever is interested and I already have a list of folks. More on this today. We'll of course have more PCBs in round 2. I don't want to make any money on this--so I would just like to charge the PCB cost and shipping.
2. Write a post on powering it up and adjusting the operating points.
3. Revise the PCB for next round.
Other than that, the PS and regulator works fine. I would also recommend not using the jumpers that shunt the zener references. I added this feature mainly for me to play around with the supply voltage, but it is not needed for a build. You can just put the zener stack in and don't wire any jumpers. The zener stack should total to about 100-130V. I'll add this to the notes on the schematic.
What's left to do:
1. Get PCBs out to those who are interested. I received six in the original order and I'd like to keep two. So, that leaves four extra. I'm going to ship these for free to whoever is interested and I already have a list of folks. More on this today. We'll of course have more PCBs in round 2. I don't want to make any money on this--so I would just like to charge the PCB cost and shipping.
2. Write a post on powering it up and adjusting the operating points.
3. Revise the PCB for next round.
Posting the latest schematic for reference. I have emailed four folks who reached out first about the extra PCBs. I know there are others who want a PCB as well. Hope to get a new batch soon.
Last edited:
Yes, it does. No problem at all. And in fact with the caps and regulator in place, the voltage comes up slowly. In the next batch, we might be able to ditch it altogether. Still testing that possibility. But the PCB works, no turn on or turn off thump.
I have tried various sizes, but right now been playing around with a 10uF, though it won't fit on the PCB. I found that the sonic signature is the same between a 56 uF and a 10 uF Wima DC link. So, perhaps a smaller size that would fit would work just as well, though I have not tested this theory.
ra7,
Thank you.
Last question. What is the lead spacing for C104/204? I think that is the last part I need to add to my shopping cart.
V/R,
Brad
Thank you.
Last question. What is the lead spacing for C104/204? I think that is the last part I need to add to my shopping cart.
V/R,
Brad
Here's the full BOM as an output from KiCAD. You can open this in Excel as a tab delimited file.
C104/204 are D10.0mm and P5.0mm.
Use this for the footprints. Some values have changed. See below.
The main values that have changed are:
Rsch: 10k
Rg: 1k (reverted back to 10k/1k because with Rg at 10k, the output impedance is too high for some amps, and it sounds just about the same or better with the 10k/1k combo)
RV102: with the lower values above, this pot should be 10k. The 100k pot will also work, but it will sweep the range with fewer turns.
It's not like the preamp won't work with 100k/10k and you can certainly try it.
I would get Kiwame for the above two resistors. Those two and C102 have a significant influence on the sound.
Other notes:
Rlim: This is a fail safe in case the wiper on RV101 fails. Right now, I put 10R in there, but you could go higher to limit the current further. Higher value = lower current limit. You could also replace the pot with a jumper and put the correct value in Rlim for the CCS current. But I use a pot for flexibility with different FETs. Probably a limit of 50 mA would be enough and I will test that and post what value of Rlim would get you 50 mA. Of course, if you are using TO-92 FETs for gain, they will be toast at 50 mA. But it would protect the rest of the circuitry and of course, FETs are cheap.
C101 and 103 can be smaller than 10 uF, but probably not smaller than 1 uF. It depends on what comes before and after this circuit.
C102: There is space for an optional bypass cap there with a pitch of 37.5mm (let me confirm that), but it is not very wide and so, only small values of film caps (Wima DC link) will fit there.
Brad, I will update the BOM over the weekend and add these and other notes to individual components.
C104/204 are D10.0mm and P5.0mm.
Use this for the footprints. Some values have changed. See below.
The main values that have changed are:
Rsch: 10k
Rg: 1k (reverted back to 10k/1k because with Rg at 10k, the output impedance is too high for some amps, and it sounds just about the same or better with the 10k/1k combo)
RV102: with the lower values above, this pot should be 10k. The 100k pot will also work, but it will sweep the range with fewer turns.
It's not like the preamp won't work with 100k/10k and you can certainly try it.
I would get Kiwame for the above two resistors. Those two and C102 have a significant influence on the sound.
Other notes:
Rlim: This is a fail safe in case the wiper on RV101 fails. Right now, I put 10R in there, but you could go higher to limit the current further. Higher value = lower current limit. You could also replace the pot with a jumper and put the correct value in Rlim for the CCS current. But I use a pot for flexibility with different FETs. Probably a limit of 50 mA would be enough and I will test that and post what value of Rlim would get you 50 mA. Of course, if you are using TO-92 FETs for gain, they will be toast at 50 mA. But it would protect the rest of the circuitry and of course, FETs are cheap.
C101 and 103 can be smaller than 10 uF, but probably not smaller than 1 uF. It depends on what comes before and after this circuit.
C102: There is space for an optional bypass cap there with a pitch of 37.5mm (let me confirm that), but it is not very wide and so, only small values of film caps (Wima DC link) will fit there.
Brad, I will update the BOM over the weekend and add these and other notes to individual components.
Yes, assuming 25 mA as max current draw, 100R would dissipate 0.0625 W. So, 0.6 W resistor should work.
STQ1NK80ZR-AP is just as good.
Friends, I made progress on preparing the packages of PCBs and FETs but didn't get them out yet. Hope to do so tomorrow or Tuesday.
Also been working on the BOM, but not quite finished yet. Progress has been slow with the usual family obligations coming first.
BRN, lead spacing on C102 bypass is 27.5 mm.
Also been working on the BOM, but not quite finished yet. Progress has been slow with the usual family obligations coming first.
BRN, lead spacing on C102 bypass is 27.5 mm.
Last edited:
Yes, Steve I got you covered. It will have to be the next batch. Already got the four extras assigned.