50W Amp TO-264 (and TO-247/TO-3P) version:
AEM6000 Based 50W Amp Rev 2.zip - Google Drive
I've included all KiCad design files, to encourage you lot to try KiCad, plus a BOM in numbers format as well as excel format, plus interactive bom, plus gerbers.
Bear in mind I'm still building this, though the differences from the previous 50W version are very small so I expect it to work without issues. If you want to build it with double-die MOSFETs, you'll have to wait for me to prototype and sort the compensation out - or you could sneak a peek at my 100W amp for where I will start with compensation caps etc.
AEM6000 Based 50W Amp Rev 2.zip - Google Drive
I've included all KiCad design files, to encourage you lot to try KiCad, plus a BOM in numbers format as well as excel format, plus interactive bom, plus gerbers.
Bear in mind I'm still building this, though the differences from the previous 50W version are very small so I expect it to work without issues. If you want to build it with double-die MOSFETs, you'll have to wait for me to prototype and sort the compensation out - or you could sneak a peek at my 100W amp for where I will start with compensation caps etc.
Hi,
building this amp for my purposes, i want to change some values of parts.
1st
- change of input impedance to 47kOhm
R1 will be 47k
C1 will be 1µF foil capacitor
2nd
change of input sensitivity to 1Veff at a supply power of +/- 34V DC
- is 22k for R48 (R49 1k) or 680R for R49 (with R48 15k) a doable way to reach this.
All used speakers here have an impedance of 4 ohms, so there will be enough output power, even with reduced supply power.
BR
Günni
building this amp for my purposes, i want to change some values of parts.
1st
- change of input impedance to 47kOhm
R1 will be 47k
C1 will be 1µF foil capacitor
2nd
change of input sensitivity to 1Veff at a supply power of +/- 34V DC
- is 22k for R48 (R49 1k) or 680R for R49 (with R48 15k) a doable way to reach this.
All used speakers here have an impedance of 4 ohms, so there will be enough output power, even with reduced supply power.
BR
Günni
I can't see any issues. I designed the newer boards (Rev 2) to take a 7.2 x 3.5 cap for C1. Wima do a 1µ MKS2 film in this size. The older boards will accept 13 x 4 caps. Again, there's a 1µ Wima MKS4 in this size.
Also, increasing the gain is fine, though you'll take a small distortion hit.
Also, increasing the gain is fine, though you'll take a small distortion hit.
The Jfets arrived today so I am starting my build soon. In order to justify the shipping I bought extra SST404 Fets so if anyone is interested I will offer two sets of boards with SST404 and MMBTA56/06 installed.
The choice is either the original TO-3P version or the newer REV2 with TO-247 packages.
I have two sets of the TO-247 boards but only one set of the TO-3P boards. The PCB are high quality Enig finish in black.
The fets are £7.50 each, the PCB's £4. The BJT's about £1
Each set of two PCB with Fets and BJT's installed - £25 GBP plus shipping
One spare set of REV2 boards - £8 plus shipping
The choice is either the original TO-3P version or the newer REV2 with TO-247 packages.
I have two sets of the TO-247 boards but only one set of the TO-3P boards. The PCB are high quality Enig finish in black.
The fets are £7.50 each, the PCB's £4. The BJT's about £1
Each set of two PCB with Fets and BJT's installed - £25 GBP plus shipping
One spare set of REV2 boards - £8 plus shipping
Attachments
Nice weekend project. MELF's are the way forward
Nice & tidy work! Did you solder by hand or in a reflow oven ?
Thank you Jacques and Suzy. Done by hand very carefully, these are nice boards to build. I'm sure there is some therapeutic value to SMD work, especially with MELF packages.
Can't believe a 50W amp is the size of a credit card
My proposed build will likely be an integrated with a muses volume and two Sjostrom diamond buffers which I have found to be a great combination.
Trying out a layout below with CRCRC around the toroid and Mark Johnsons Cap multiplier before the amps. Aiming for a compact build.
Can't believe a 50W amp is the size of a credit card
My proposed build will likely be an integrated with a muses volume and two Sjostrom diamond buffers which I have found to be a great combination.
Trying out a layout below with CRCRC around the toroid and Mark Johnsons Cap multiplier before the amps. Aiming for a compact build.
Attachments
I really enjoy working with MELF parts. With the right tweezers they’re no harder than 1206, but they have lovely big solid end caps.
Have you tried the rev 2 PCB yet? I subtly revised the MELF footprint on that rev to space the pads slightly further apart, with a teensy bit more copper, so they’re really easy to hand solder.
Oh, and I’ll let you in on a secret. The rev 2 board should do 100W with dual-die parts, despite the silkscreen claiming otherwise.
Have you tried the rev 2 PCB yet? I subtly revised the MELF footprint on that rev to space the pads slightly further apart, with a teensy bit more copper, so they’re really easy to hand solder.
Oh, and I’ll let you in on a secret. The rev 2 board should do 100W with dual-die parts, despite the silkscreen claiming otherwise.
At last the vital ingredient has arrived, a 125VA custom 'O' core from James Transformers. the first O core I have used, they have the benefits of both R-core and toroidal transformers apparently. Slightly wider and shorter than the same spec toroidal.
James makes very nice transformers, his prices are decent.
Should be able to get some sound out of this in the next couple of days
James makes very nice transformers, his prices are decent.
Should be able to get some sound out of this in the next couple of days
Attachments
Hey everyone!
So first of all, thank you for designing and publishing this awesome looking (and apparently also sounding ;-) ) amp, Suzy!
I would like to be powering a stereo pair with a toroidal transformer I still have on hand with two 24V secondary windings, which should come out to something like +/- 32Vdc rail voltage.
I will be powering a pair of 4 Ohm bookshelf style speakers with it.
In the construction notes you have published, I saw that a change in load resistance R6 and R7 might be recommended for different rail voltages/load impedances.
Would you or someone else be able to give me a short explanation on how to calculate this and/or make a recommendation on the resistance I should use in my case.
Thanks a lot
So first of all, thank you for designing and publishing this awesome looking (and apparently also sounding ;-) ) amp, Suzy!
I would like to be powering a stereo pair with a toroidal transformer I still have on hand with two 24V secondary windings, which should come out to something like +/- 32Vdc rail voltage.
I will be powering a pair of 4 Ohm bookshelf style speakers with it.
In the construction notes you have published, I saw that a change in load resistance R6 and R7 might be recommended for different rail voltages/load impedances.
Would you or someone else be able to give me a short explanation on how to calculate this and/or make a recommendation on the resistance I should use in my case.
Thanks a lot
Easy peasy! It’s biased to have ~800μA running down each side of the input stage. The cascode transistors are biased to put 11V across the input JFETs by the zener. The voltage across the load resistors R6 and R7 is 800μA x 18K = 14V, so at 40V supply, the collector of the cascade transistors is at 26V, giving me 15V across the cascode transistors. Everything is forward biased, so it all works.
If you knock the supply down to 32V and don’t change anything, you still get 7V or thereabouts across the cascodes, which should be fine. It’s pretty conservative in this respect. You could reduce R6 and R7 to say 15K, which will give you more headroom, but will lower first stage voltage gain by a little, with a small THD penalty.
I’d leave it as is.
If you knock the supply down to 32V and don’t change anything, you still get 7V or thereabouts across the cascodes, which should be fine. It’s pretty conservative in this respect. You could reduce R6 and R7 to say 15K, which will give you more headroom, but will lower first stage voltage gain by a little, with a small THD penalty.
I’d leave it as is.