I just built up the MB and the Vcc power rail test was successful. I am missing some electrolytic caps for the CRCRC but otherwise everything went together pretty easily. I wish the input and output bypass film caps allowed 5mm pitch Wima MKS because I have bags of those. I had to bend the pins 90deg and "surface mount" them on the outputs. Otherwise, and excellent layout. Of course, if you bought the items in the BOM, they would all fit. This test build is using strictly stuff I have on my bench without ordering anything. The signal input/output connector is something I don't have at present, but the wiring configuration is easy to understand: inputs on the two inside pins and corresponding outputs to the outside pins in the lower row, the top row are all GND. I am using an RK09 Alps pot from my pocket amp and attached a short piece of vinyl tubing to allow the shaft to be operated from the front.
Here is a 44.5v DC step up putting out about 40.5v after the cap Mx and CRCRC. I am using a non-ideal 470uF cap on the cap Mx so the slow ramp up takes about 60 seconds to reach full voltage. That's actually nice from standpoint of preventing thumps and other transients. 220uF or 100uF would be fine. One thing is for sure, this PSU will be dead quiet with a flat noise background down at the measurement limit. I have typically seen -130dB and flat.
So I will build up the amp core daughterboards next when I have some more time. All for now...
Here is a 44.5v DC step up putting out about 40.5v after the cap Mx and CRCRC. I am using a non-ideal 470uF cap on the cap Mx so the slow ramp up takes about 60 seconds to reach full voltage. That's actually nice from standpoint of preventing thumps and other transients. 220uF or 100uF would be fine. One thing is for sure, this PSU will be dead quiet with a flat noise background down at the measurement limit. I have typically seen -130dB and flat.
So I will build up the amp core daughterboards next when I have some more time. All for now...
Attachments
Last edited:
GB MB & TH PCBs are Singing
I built up the TH daughterboard using a 12k carbon film feedback resistor and 10R LTP degeneration resistors. Using a BSP129 in place of DN2540. Only one channel built so far. This is with 45V Vcc.
FFT for 8Vpp into 7kohms (THD is 0.00071%):
FFT for 20Vpp into 7kohms (THD is 0.0016%):
If you increase Vcc to 48V, you get THD of 0.0001%, but H2 and H3 become closer together. R9 is currently 100k, reducing that to 68k will increase H2 relative to H3.
I built up the TH daughterboard using a 12k carbon film feedback resistor and 10R LTP degeneration resistors. Using a BSP129 in place of DN2540. Only one channel built so far. This is with 45V Vcc.
FFT for 8Vpp into 7kohms (THD is 0.00071%):
FFT for 20Vpp into 7kohms (THD is 0.0016%):
If you increase Vcc to 48V, you get THD of 0.0001%, but H2 and H3 become closer together. R9 is currently 100k, reducing that to 68k will increase H2 relative to H3.
Attachments
Last edited:
The MELF GB is still on. Yes, I will ship to Canada address. Is this the address on your PayPal? That’s where I ship to so that he order can be closed out.
Hmm... the topology is sort of similar on the LTP with diode but JFET Input, NPN on the right LTP leg but wired differently. Also has a push pull output stage vs SE Class A output on this one. LF353 spec says THD is 0.02% - quite a bit higher than his amp.
Attachments
Last edited:
Paypal address
Hi X,
The Canada address is the PayPal shipping address. I know you will ship the board sets to Canada.
I was trying to confirm that JPS64 would be invoicing for the MELF GB, and would need the correct shipping/mailing address. Sounds like everything is OK.
K
Hi X,
The Canada address is the PayPal shipping address. I know you will ship the board sets to Canada.
I was trying to confirm that JPS64 would be invoicing for the MELF GB, and would need the correct shipping/mailing address. Sounds like everything is OK.
K
Hi X,
I just checked my Paypal account. There is a billing address and a primary address, both of them Mexico. When JPS64 invoices for the MELFS, I will PM him and supply him the Canada shipping address. Sorry for the confusion.
Here are some suggested alternate resistor values to let you tune H2/H3 and overall level of THD, and amp gain to your taste:
All values on original schematic are good, just order extra values in red to allow adjustment.
As starting values I would recommend R5/R6=33R, R10=6k8, and R9=68k. Have a listen and go from there. Mostly increase/decrease R10 to get the gain you need. With a normal line level source, you will need the 12k to hit the full 40Vpp, but if you have a more powerful DAC with full +/-15v opamp output, perhaps only 6k8 is needed. For use as a line-level amp (not for driving unity gain buffers), 3k3 is plenty as a preamp.
R5/R6 lower values give lower overall THD but higher H3 relative to H2. Higher R5/R6 gives overall higher THD.
Lower value of R9 gives more H2 relative to H3 but increases overall THD.
Data shown above is for R5/R6=10R, R9=100k, R10=12k.
All values on original schematic are good, just order extra values in red to allow adjustment.
As starting values I would recommend R5/R6=33R, R10=6k8, and R9=68k. Have a listen and go from there. Mostly increase/decrease R10 to get the gain you need. With a normal line level source, you will need the 12k to hit the full 40Vpp, but if you have a more powerful DAC with full +/-15v opamp output, perhaps only 6k8 is needed. For use as a line-level amp (not for driving unity gain buffers), 3k3 is plenty as a preamp.
R5/R6 lower values give lower overall THD but higher H3 relative to H2. Higher R5/R6 gives overall higher THD.
Lower value of R9 gives more H2 relative to H3 but increases overall THD.
Data shown above is for R5/R6=10R, R9=100k, R10=12k.
Last edited:
GB Status update:
I will be sending out GB1 boards mid next week as I will be on travel for a few days. TH daughterboard has been verified to work, but TH so far has not yet as I don't have the MELFs. However, I am pretty confident it will work as layout is similar to TH one except in SMT.
I will be sending out GB1 boards mid next week as I will be on travel for a few days. TH daughterboard has been verified to work, but TH so far has not yet as I don't have the MELFs. However, I am pretty confident it will work as layout is similar to TH one except in SMT.
Last edited: