OK! All wired up and it works fine I think. I forgot to tie the ground side of the transformer primaries at first and was puzzled at the thin sound…
The frequency response is really flat, I put a bit of a trebly curve on my cell phone (yes I tested it with MP3s on my cell phone…) to make it sound more like how I set my stereo. Overall it's really, really clear at low to mid levels. It definitely gets a little mushy just past "normal" volume… I would like a tiny bit more headroom. But then again my DT770 are 96db/mW and there seem to be more sensitive phones out there… they are also the discontinued 600 ohm versions and I guess a set of 250 ohm on the 250-600 ohm taps would be louder too.
Also just ZERO noise. Dead quiet so far.
It's totally going in a box in the next couple weeks.
My ECC99s seems to only be running at 13mA (3.9V/300 ohm).. should I try to jack that up a bit?
Brian
The frequency response is really flat, I put a bit of a trebly curve on my cell phone (yes I tested it with MP3s on my cell phone…) to make it sound more like how I set my stereo. Overall it's really, really clear at low to mid levels. It definitely gets a little mushy just past "normal" volume… I would like a tiny bit more headroom. But then again my DT770 are 96db/mW and there seem to be more sensitive phones out there… they are also the discontinued 600 ohm versions and I guess a set of 250 ohm on the 250-600 ohm taps would be louder too.
Also just ZERO noise. Dead quiet so far.
It's totally going in a box in the next couple weeks.
My ECC99s seems to only be running at 13mA (3.9V/300 ohm).. should I try to jack that up a bit?
Brian
My rough rule of thumb is to ensure there is enough headroom to reach 117dB SPL which is very loud by any standards. This gives enough headroom for transients. Your phones have a 96dB/mW sensitivity so the reach 117dB SPL they need a level of +21dBu (just over 100mW of power) which is about 8.7V rms. You should have the transformer secondary wired for the smallest ratio of 4:1 so the primary side voltage needs to reach close to 35V rms which is 98V peak to peak. The centre point of the SRPP therefore needs to swing by plus or minus 50V. With a 300V supply this should present no problem.
The 600 ohm headphones reflected to the primary looks like 9600 ohms. With 35V rms across it that is a load current of just over 3.6mA rms or 5.2mA peak. This implies 13mA of quiescent current is plenty. With a 300V supply, the dissipation per triode is 150 x 13 mW = 1.85W. It is not clear from the datasheet if the limiting power dissipation of 5W is for the whole tube or per triode. The 6CG7, which is very much the same size, has a 3.5W per triode dissipation spec so let's assume the ECC99 is 2.5W per triode. To raise it a bit higher the simplest way is to reduce the cathode resistor. I suggest you try 250 ohms. You can afford to go up to just over 16mA at 150V before reaching 2.5W dissipation. The extra current is not really required but the reduction in plate resistance should improve the drive capability.
It is unfortunate that the windings of the Sowter transformer are a bit of a compromise in order to span a very wide range of headphone impedances with the higher impedance end of each range resulting in the lesser performance. For driving 600 ohm phones a ratio of 2:1 would probably be better. If you want to try a low cost alternative then I suggest you get a couple of the incredibly cheap Edcor XSM2.4K/600:
https://www.edcorusa.com/xsm24k-600
They are not in the same class as the Sowter but for the price they are probably worth a try.
This headphone amp design morphed into a general purpose line amplifier (in the pro audio rather than the hi-fi sense) with variable gain. It is more of less the same topology but the output tubes are 6922 (E88CC) run at about 5mA. With a 2K4:600 transformer this will easily output +20dBu into a 600 ohm load. I don't use Edcors with is but the much better but still good value from money Carnhill VTB2291 transformers. I mention this just to illustrate that the idle current of 13mA you have should be more than adequate for a 600 ohm load.
Cheers
Ian
Ian,
Thanks as always for taking the time to break it all down for me. I played around today for a little while... I didn't go so far at to swap out the 300 ohm Rs for 250, but I did try led bias on the lower triodes. I have a collection of leds but none with a forward drop of more than 3.9vDC. current shot up 20mA. I can't be sure anything sounded better but surely no worse. Have you ever tried led bias on an srpp?
I think the distortion might be from my headphones actually, for a few reasons. Firstly, the onset of audible distortion is lower than 8vpp; secondly there is no visible clipping at voltages way higher than this (sine wave); and thirdly I pulled out a set of 63 ohm Sony MDR-7506 (1000mW handling) and they seems to handle the same input levels from my ipod with no distortion that I could hear.... though they are much more sensitive and they were plugged into the higher impedance tap so they were so loud I could only listen for a few seconds before pulling them off. I'll find some time to rewire the tap and so a proper comparative loudness test soon. In summary though I suspect my dt770 are either truly hitting their limit or their limit has come down due to age and use. Bear in mind this all at levels higher than I'll typically listen to, but I'm trying to test the limits of the amp as much as anything else and I might need to test other phones to be sure I'm hearing the amp and not the phones.
Thanks as always for taking the time to break it all down for me. I played around today for a little while... I didn't go so far at to swap out the 300 ohm Rs for 250, but I did try led bias on the lower triodes. I have a collection of leds but none with a forward drop of more than 3.9vDC. current shot up 20mA. I can't be sure anything sounded better but surely no worse. Have you ever tried led bias on an srpp?
I think the distortion might be from my headphones actually, for a few reasons. Firstly, the onset of audible distortion is lower than 8vpp; secondly there is no visible clipping at voltages way higher than this (sine wave); and thirdly I pulled out a set of 63 ohm Sony MDR-7506 (1000mW handling) and they seems to handle the same input levels from my ipod with no distortion that I could hear.... though they are much more sensitive and they were plugged into the higher impedance tap so they were so loud I could only listen for a few seconds before pulling them off. I'll find some time to rewire the tap and so a proper comparative loudness test soon. In summary though I suspect my dt770 are either truly hitting their limit or their limit has come down due to age and use. Bear in mind this all at levels higher than I'll typically listen to, but I'm trying to test the limits of the amp as much as anything else and I might need to test other phones to be sure I'm hearing the amp and not the phones.
Brian,
Perhaps the easiest way to test the limits of the amp is to load it with 600 ohms and crank up the input with a sine wave and look at the output on a scope. At some point one half of the waveform will begin to clip because the circuit is not perfectly balanced (they never are). Note this level and then back off by about 3dB. The output should now be comfortably close to the top of the 'linear range'. If you now have about 8V rms or more then all is fine.
I have never tried LED or diode biasing an SRPP and I'm not sure it would be beneficial as it provides a sort of fixed bias. I prefer some NFB at dc to stabilise the operating point.
Cheers
Ian
Perhaps the easiest way to test the limits of the amp is to load it with 600 ohms and crank up the input with a sine wave and look at the output on a scope. At some point one half of the waveform will begin to clip because the circuit is not perfectly balanced (they never are). Note this level and then back off by about 3dB. The output should now be comfortably close to the top of the 'linear range'. If you now have about 8V rms or more then all is fine.
I have never tried LED or diode biasing an SRPP and I'm not sure it would be beneficial as it provides a sort of fixed bias. I prefer some NFB at dc to stabilise the operating point.
Cheers
Ian
Project not dead - Update
Hi all,
I'm still working on this. I designed a chassis and found a place to laser cut and fold it up…
Im test fitting everything and it looks good so far. The box is a bit of a beast though at 12"x8"x3". I'll be painting it this week and bring to <finally> put it together next weekend.
Brian
Hi all,
I'm still working on this. I designed a chassis and found a place to laser cut and fold it up…
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
Im test fitting everything and it looks good so far. The box is a bit of a beast though at 12"x8"x3". I'll be painting it this week and bring to <finally> put it together next weekend.
Brian
Last edited:
And a little more done today:
I polished off any oxide from the areas where the screws will pull the box closed and applied a hand irridite solution (to help maintain conductivity with age), then masked those spots off with kapton tape and asked a friend at work to powder coat the enclosure… came out great.
I dismantled the circuit that was thrown together on the work bench and started transposing it into the enclosure. I got sleepy and had to stop before I scratched something.
Shockingly everything fits without too much fuss. A couple holes needed to be drilled out. I thought that the enclosure may have been too big but now I think it's perfect.
I polished off any oxide from the areas where the screws will pull the box closed and applied a hand irridite solution (to help maintain conductivity with age), then masked those spots off with kapton tape and asked a friend at work to powder coat the enclosure… came out great.
I dismantled the circuit that was thrown together on the work bench and started transposing it into the enclosure. I got sleepy and had to stop before I scratched something.
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
An externally hosted image should be here but it was not working when we last tested it.
[/URL]
Shockingly everything fits without too much fuss. A couple holes needed to be drilled out. I thought that the enclosure may have been too big but now I think it's perfect.
OK, 99% complete. The only thing not done is the impedance switch, but I'm only using high impedance headphones right now anyway. I ended up pulling the whole voltage regulator section in favor of some larger capacitors and 2 series dropping resistors… Everything work fine in a tangled mess on my work bench, as soon as it was built up properly the regulator oscillated wildly. I'm now going bridge rectifier > 250uF cap > 1k resistor > 100uF > 1k > 100uF and straight to the ecc99 plates… 290vDC under load. There are some 4k7 series droppers/ 8uF filer caps feeding the 12AX7. No noise, slight hiss near max volume.
Currently in service next to my couch where I'm finally putting some hours on it.
Brian
Currently in service next to my couch where I'm finally putting some hours on it.
Brian
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.