I am building a MOSFET based headphone amp based on this design;
Szekeres headphone amp
The circuit requires a regulated PSU. Would a full wave rectified circuit based around a 7812 regulator (12v) with 2200uF on the rail, using a 18-0-18 (55VA) transformer fit the bill. or would I need to be a little more adventuress?
regards,
Chris.
Szekeres headphone amp
The circuit requires a regulated PSU. Would a full wave rectified circuit based around a 7812 regulator (12v) with 2200uF on the rail, using a 18-0-18 (55VA) transformer fit the bill. or would I need to be a little more adventuress?
regards,
Chris.
Well, the schematic says 9-15 V regulated, however as I see it these are rather conservative numbers.
Since you want to use a 18V toroid, you will get about 25V after the rectifier. Instead of burning half of it uselessly in the regulator, I would recommend going for atleast a 15V variant, 7x15.
However I see no reason why one should not go further up with the supply, 18V seems nice, use 7x18 regulators.
Of course power dissipation in the follower will increase, but 2.5W is not that much, and that would go up to like 3.6W - use suitable heatsink.
With 200mA current and 2200uF caps you end up with about 1V ripple, so 18V rails still give enough headroom.
However please check my numbers, they're just quick estimates
Have fun, Hannes
Since you want to use a 18V toroid, you will get about 25V after the rectifier. Instead of burning half of it uselessly in the regulator, I would recommend going for atleast a 15V variant, 7x15.
However I see no reason why one should not go further up with the supply, 18V seems nice, use 7x18 regulators.
Of course power dissipation in the follower will increase, but 2.5W is not that much, and that would go up to like 3.6W - use suitable heatsink.
With 200mA current and 2200uF caps you end up with about 1V ripple, so 18V rails still give enough headroom.
However please check my numbers, they're just quick estimates
Have fun, Hannes
I'd shoot for the lower end of the range. Without voltage gain all this circuit does is ensure that there is plenty of current for low impedance phones.
There might be some linearity benefit to higher rails. Nelson Pass mentioned that mosfets like to see 20V across them.
I assume that the proposed transformer is already in your parts bin. Since you have voltage to burn, consider feeding your regulators with a pi filter (C-R-C or C-L-C) or even R-C-R-C. This will reduce the ripple feeding the regulator, giving you quieter output. I'd use 1 ohm 5W resistors in the filter.
Your 7812/7815 would probably work just fine. Some prefer the sound of discrete regulators, but for this application I doubt that you'll hear a difference.
You can substitute a larger cap for C2. Bypass it with a 1 uf film cap and a 100 nf film cap. C1 also ought to be a film cap, larger is OK.
You might try the Pass bias philosophy - more is better. Swap out R2 for lower values until you find the most that your heat sinks can stand. Watch the dissipation in R4 - consider an aluminum one mounted to the case or a heat sink. Probably overkill, this burning 40W for a few mW of output...
Have fun.
There might be some linearity benefit to higher rails. Nelson Pass mentioned that mosfets like to see 20V across them.
I assume that the proposed transformer is already in your parts bin. Since you have voltage to burn, consider feeding your regulators with a pi filter (C-R-C or C-L-C) or even R-C-R-C. This will reduce the ripple feeding the regulator, giving you quieter output. I'd use 1 ohm 5W resistors in the filter.
Your 7812/7815 would probably work just fine. Some prefer the sound of discrete regulators, but for this application I doubt that you'll hear a difference.
You can substitute a larger cap for C2. Bypass it with a 1 uf film cap and a 100 nf film cap. C1 also ought to be a film cap, larger is OK.
You might try the Pass bias philosophy - more is better. Swap out R2 for lower values until you find the most that your heat sinks can stand. Watch the dissipation in R4 - consider an aluminum one mounted to the case or a heat sink. Probably overkill, this burning 40W for a few mW of output...
Have fun.
Bob,
200mA bias current for headphones is already huge. Please do not forget that headphones are about a factor 1000 more sensitive than speakers. With 200mA you can already drive horns.
Further I would avoid a CRC-filter infront of the regulator.
Reason is that standard regulators work incredibly well at this 100Hz ripple frequency and a CRC-filter will only increase power supply impedance.
If he wants top performance he can use a CRC-filter after the regulator to filter noise. The additional voltage drop is not a problem in ClassA-circuits.
As you already mentioned, mosfets want to see higher rails, as this lowers their large capacitances and as he already has an oversized transformer his best option would be to use it for something reasonable.
But of course, everbody is welcome to do his own experiments!
Have fun, Hannes
200mA bias current for headphones is already huge. Please do not forget that headphones are about a factor 1000 more sensitive than speakers. With 200mA you can already drive horns.
Further I would avoid a CRC-filter infront of the regulator.
Reason is that standard regulators work incredibly well at this 100Hz ripple frequency and a CRC-filter will only increase power supply impedance.
If he wants top performance he can use a CRC-filter after the regulator to filter noise. The additional voltage drop is not a problem in ClassA-circuits.
As you already mentioned, mosfets want to see higher rails, as this lowers their large capacitances and as he already has an oversized transformer his best option would be to use it for something reasonable.
But of course, everbody is welcome to do his own experiments!
Have fun, Hannes
I have completed the amp modules (1 board per channel) using a modified szerkeres design. The design utilises a LM317 voltage regulator as a CCS runing at 250mA. I have used 470uF Silmic II's, bypassed with Wima 1uf film caps as the output capacitors. I have tested them using a Maplin 13.8V/3A regulated PSU, (although I will build a PSU eventualy), and a 100k audio taper pot, which is a little quiet (I have a 50K Alps Blue ready to use). The amp modules sound very detailed and clear, if a little bright. The output caps will need a long burn in I think.
Is there a good way to connect the amps to the PSU (i.e a good grounding scheme), as when I connected both amps as a stereo pair, sound was quiet and distorted?
Regards,
Chris.
Is there a good way to connect the amps to the PSU (i.e a good grounding scheme), as when I connected both amps as a stereo pair, sound was quiet and distorted?
Regards,
Chris.
I found this thread a bit confusing, since I hadn't looked at the amp schematic, and I assumed from the content that you needed a dual rail supply, then I looked at the schematic and it's single-rail?? Anyway, if you're running it from a lab psu...
Spider the power connections, don't daisy-chain them and don't form a loop, it's not a ring-main. Keep the wires short and of about equal length, but this should not be critical in a preliminary test.
Both amps worked independently. It doesn't sound like a grounding issue, too unsubtle. You need to look elsewhere for the problem. It's probably a simple wiring error or solder string.
Check the power supply isn't current-limiting...
w
The absolute value of the pot won't make any difference to the volume.
Spider the power connections, don't daisy-chain them and don't form a loop, it's not a ring-main. Keep the wires short and of about equal length, but this should not be critical in a preliminary test.
Both amps worked independently. It doesn't sound like a grounding issue, too unsubtle. You need to look elsewhere for the problem. It's probably a simple wiring error or solder string.
Check the power supply isn't current-limiting...
w
The absolute value of the pot won't make any difference to the volume.
Re: is there any chance ?????
I have included a diagram of the modified cct (from the original web site, project adendum page). It probably could drive 8ohm speakers, but they would have to be very sensetive, and I would have to increase the size ot the heatsinks and power supply. I would also have to change the output caps to allow for the change in impedence from the 32ohm's I have calculated for.
wakibaki,
I am using a 3A PSU. I think it should have enough power on tap, although I could measure the current consumtion if needed. The modules were connected seperately to the PSU terminals. The signal ground's were also connected seperately to the PSU ground terminal. Each amp module functions on it's own, if a little quiet. I was careful with the soldering, but I will desolder the wires and put it back together when I get a bit of spare time.
Regards,
Chris.
sakis said:
I have included a diagram of the modified cct (from the original web site, project adendum page). It probably could drive 8ohm speakers, but they would have to be very sensetive, and I would have to increase the size ot the heatsinks and power supply. I would also have to change the output caps to allow for the change in impedence from the 32ohm's I have calculated for.
wakibaki,
I am using a 3A PSU. I think it should have enough power on tap, although I could measure the current consumtion if needed. The modules were connected seperately to the PSU terminals. The signal ground's were also connected seperately to the PSU ground terminal. Each amp module functions on it's own, if a little quiet. I was careful with the soldering, but I will desolder the wires and put it back together when I get a bit of spare time.
Regards,
Chris.
Attachments
Re: Re: is there any chance ?????
Lots of lab supplies have 2 controls, voltage and current limiting.
You can set the maximum current permitted, and the PSU will reduce the output voltage to prevent it being exceeded.
I thought your Maplin one might be one of these. You could have one module working fine, connect up the second and have the current limiting kick in unnoticed.
Not very likely perhaps, but why would two working modules get quiet and distorted? Stranger things have happened at sea...
13.8/3A sounds like a CB psu tho', but that didn't register with me first time round.
w
zebra100 said:
Lots of lab supplies have 2 controls, voltage and current limiting.
You can set the maximum current permitted, and the PSU will reduce the output voltage to prevent it being exceeded.
I thought your Maplin one might be one of these. You could have one module working fine, connect up the second and have the current limiting kick in unnoticed.
Not very likely perhaps, but why would two working modules get quiet and distorted? Stranger things have happened at sea...
13.8/3A sounds like a CB psu tho', but that didn't register with me first time round.
w
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.