mwl168

I am starting a new build thread for the PSU (GR78, GR79, GRHVxxx and the main board) so all can share and exchange experience.

This whole thing is just odd - less than 24 hours after getting a call from FedEx to inform me that the "shipper" in Taiwan had sent FedEx the wrong item to ship and the correct item has not been shipped and I should contact the shipper directly, the boards showed up yesterday (I know that PCBnet has some of their boards fabricated in their Taiwan facility). As expected, these 2mm boards look and feel great! Going to start packaging boards for shipment and I should receive the GRHVxxxx bare boards next week.

The latest FedEx tracking shows our boards are "in transit" in Japan I don't know if to laugh or cry!

FedEx was supposed to deliver the re-manufactured PSU main boards yesterday. FedEx messed up the delivery (long story), I ended up calling FedEx customer support to straight things out. Wes put on hold for 40+ minutes, finally talked to a rep and confirmed address and contact info. Then, nothing - tracking info still shows "pending" due to incorrect address today. Called customer support again just now, went through the same routine. Let's see what happens next. It shouldn't be this hard!

Belated Happy Birthday Chris!

Just got these in...

What? No love for the Pioneer 8 Track player? I grew up with those things!

I also recall the terminal blocks shorting issue also had to do with too much solder being applied.

Some bending but no twisting...

The board I will be building has 2SC3381. I don't have the board in hand yet. Looking at the 2SC3381 and MPSA06 pinout, one should be able to mount the 2 MPSA06 into the 2SC3381 position without twisting the pins by positioning them in different orientation therefore not possible to bond the two MPSA06 together.

Thanks JoaMat. I am thinking of using 2 MPSA06 mounted on the same side without bounding them together. No twisting and no dancing

Kevin has mentioned that the 2SC3381 can be replaced by a pair of MPSA06 as well. Has anyone tried that? Also, why do you put one of the KSC1008 on the opposite side?

Got good news from PCBnet - they will remanufacture the PSU main boards to the spec'ed 2mm thickness. Expect to be shipped next week. PCBnet has truly been a top notch company to do business with. Their product quality and customer service have both been excellent!

yes, the technology has existed for some time but I am hoping for the technology to be more accessible (cost, size, etc.) and better usability.

There is no way for me to prove this but I am willing to bet big dollars that the reason some times the same gears of mine sound much better (or worse) than other times has to do with the quality of the AC power from the grid at the time. Would be great if someday the battery technology is such that we can power our audio gears from DC supply that's totally isolated from the AC grid.

More goodies showed up last night but with some bad news. The PSU main PCB came in 1.6mm/3oz copper instead of 2mm/3oz copper specification. There was a miscommunication that neither I or PCBnet caught until now - the order submitted specified 2mm thickness an 3oz copper but the quote was for 0.62' and 3oz copper and manufactured accordingly. I noticed this discrepancy when I took the PCBs from their wrapping to take some photos. The mini T2 boards are 2mm/3oz copper as ordered. All the PCBs look great with a nice heavy feel to them and are very good quality as can be expected from PCBnet. I have contacted PCBnet to see what can be done to resolve the issue.

Updated GR78 and GR79 BOM in the first post to correct the C5 (47 pf) part number from 1206 to 0805. The 1206 package may fit but more difficult to solder.

R5 and R6 set the bias current of the C4686A output devices. The voltage drop across them divided by their resistance is the bias current. IIRC, the position of these two resistors varies depending on the version of the board. You’ll need to trace the board to locate them.

What's supply rail voltage and the bias current of your KGSSHV? It's probably not healthy for the active devices on the amp longterm if the heatsinks are so hot that you can barely touch it. May want to consider either reduce the voltage and/or the bias current.

Look what I just found at my door step...

I believe that 2SC4686A would be good substitute for Q8-10, Q13-15? 2SC4686A is a 1200V device available at BDent and much less expensive than 2SC3675 but lower hFE. If you are willing to take the risk, 2SC4686A is also available at Utsource and the last batch I bought from Utsource tested similar to the ones I got from BDent. Be warned that Utsource sells mix of genuine and fake parts. Based on suggestions from someone very knowledgeable here, I plan to use 2SC3675 in the active batteries (need high hFE), 2SC4686A in the output current source and FJPF2145 for all other places to substitute for 2SC3675 in the T2.

Hi JoaMat: Have you considered experimenting with STP03D200 in the active batteries to replace the 2SC3675? It's a 2KV Darlington transistor with hFE > 200. This part was actually mentioned in the DIY T2 thread long time ago. It's a current production part available from Mouser, etc. and expensive (> $6 US) but still cheaper than 2SC3675 and much easier to source.

Based on my experience, 24vac and 60va transformer is the minimum for 24VDC rails to supply a pair of CFA boards. I use two 100va transformers for my dual mono, balanced CFA amp running 30VDC rails. For GRLV, I did match the 4 resistors that determine the regulated voltage. No need for 5 decimal precision though. The LT1021 voltage reference has its own tolerance as well.

The chassis you showed should work if you can fit everything in. It looks to me a similar chassis to the one I am using for mine. I posted photos of it when I had the PSU built in. Mine is a balanced amp. Maybe consider moving this discussion to the thread dedicated to the dynamic CFA build.

Remind us are you picking a chassis for a singll-ended or balanced build? PSU in the same chassis or separate chassis?