Hi,
I have just finished and tested 2 aikido amplifiers using 12AX7's in gain stage and 12AT7's in buffer.
PSU is 375V 150mA transformer, 5Z4 rectified with plenty of capacitance and a 20H choke.
Each channel is seperatly tube regulated using 2x0D3 and 2x0C3 giving 265V per amp.
Tubes used in the Aikido are 12AX7's for gain and 12AT7's for buffer.
I get about 2mV PSU junk/HF junk on output of each channel which is ok i guess.
My frequency response is from almost DC to 50Khz (-3dB).
Gain is around 15 but my tubes are dodgy and old so it might be completely different.
Will a transistor CCS before the regulators help the 2mV noise?
What specs is this amp supposed to have and what specs have people got from building it?
How much thump do you get turning it on, i get about 5V on and 3V off (enough to mash my speakers
,when i eventually hook it up ) is this normal?
Output offset is about 30mV.
I would be interested to know what success people have had with this amp as i have heard it is very popular.
long post hehe
Cheers
Craig
I have just finished and tested 2 aikido amplifiers using 12AX7's in gain stage and 12AT7's in buffer.
PSU is 375V 150mA transformer, 5Z4 rectified with plenty of capacitance and a 20H choke.
Each channel is seperatly tube regulated using 2x0D3 and 2x0C3 giving 265V per amp.
Tubes used in the Aikido are 12AX7's for gain and 12AT7's for buffer.
I get about 2mV PSU junk/HF junk on output of each channel which is ok i guess.
My frequency response is from almost DC to 50Khz (-3dB).
Gain is around 15 but my tubes are dodgy and old so it might be completely different.
Will a transistor CCS before the regulators help the 2mV noise?
What specs is this amp supposed to have and what specs have people got from building it?
How much thump do you get turning it on, i get about 5V on and 3V off (enough to mash my speakers
,when i eventually hook it up ) is this normal?Output offset is about 30mV.
I would be interested to know what success people have had with this amp as i have heard it is very popular.
long post hehe
Cheers
Craig
Craig - I'm no help w/measurements for my unit (which is 6SN7/6SN7), but in any case I'd expect that there may be something up with yours: gain should be more like 50 than 15.
Subjectively I'm very happy with mine, no plans to change it anytime soon at all.
Re: turn-on, did you use the "safety" voltage dividers Broskie added in a later blog? They should help with turn-on (although those voltage regs do turn-on quite suddenly, that might be the issue.)
Best,
-j
Subjectively I'm very happy with mine, no plans to change it anytime soon at all.
Re: turn-on, did you use the "safety" voltage dividers Broskie added in a later blog? They should help with turn-on (although those voltage regs do turn-on quite suddenly, that might be the issue.)
Best,
-j
J Epstein, yea i used the dividers.
I might recheck the gain, that is a massive difference, although i dont know the gain of the tubes you used.
I havnt actually listened to mine yet (lackng my power amp for a bit) but i hope its as good as some people have said.
i might create a mute using Jfet's to short the o/p ofter the capacitor for a few seconds at turn on and maybe at turnoff too.
Craig
I might recheck the gain, that is a massive difference, although i dont know the gain of the tubes you used.
I havnt actually listened to mine yet (lackng my power amp for a bit) but i hope its as good as some people have said.
i might create a mute using Jfet's to short the o/p ofter the capacitor for a few seconds at turn on and maybe at turnoff too.
Craig
mine is quiet turning on and off
no noise i can hear at output either
but i'm just using two 20H chokes
choke input with two 470uf caps
and a 1k>470uf RC stage to drop the voltage
i have run mine with 12ax7 > 12at7
that had way too much gain
but sounded ok
though not as good as with 12au7 CF
i'm now using 6n1p > 6n6p and it is wonderful
dave dove
no noise i can hear at output either
but i'm just using two 20H chokes
choke input with two 470uf caps
and a 1k>470uf RC stage to drop the voltage
i have run mine with 12ax7 > 12at7
that had way too much gain
but sounded ok
though not as good as with 12au7 CF
i'm now using 6n1p > 6n6p and it is wonderful
dave dove
Just a few comments regarding Aikido power supply...
First, I recommend using a CCS in place of a resistor feeding the shunt regulators, it will help reduce noise and grunge even though the Aikido supply is fairly good in doing this intrensically.
Regarding a turnon thump, I would strongly recommend installing a mute switch or time delay relay that will ground the outputs until the regulators light off. That being said, the sound is really just annoying, it doesn't do any damage, at least in my system.
Also, I have found that power supply capacitor quality and size is clearly audible, even with shunt regulation with CCSs (I am using a similar arrangement with series 0A2s for a 300V B+). I would recommend using as small value as you can get away with without comprimising the hum or hash on the outputs). If using electrolytic caps, bypass them with signal quality film or oil caps.
That being said, the Aikido circuit is capable of very low level resolution, the naturally low PSRR means that meticulous attention to PS quality will reap great benefits in resolution.
Just one comment to Dave Dove, regarding the problem I was having a few weeks ago concerning excessive noise that I thought was attributed to the safety resistor divider, I subsequently found that the 12FQ7s I was using at the time had a problem with poor heater insulation that was producing excessive buzz noticable through my 105dB efficient horn speakers. Using a different tube solved the problem. Thanks for all the thought and effort in your comments to me on the problem.
David
First, I recommend using a CCS in place of a resistor feeding the shunt regulators, it will help reduce noise and grunge even though the Aikido supply is fairly good in doing this intrensically.
Regarding a turnon thump, I would strongly recommend installing a mute switch or time delay relay that will ground the outputs until the regulators light off. That being said, the sound is really just annoying, it doesn't do any damage, at least in my system.
Also, I have found that power supply capacitor quality and size is clearly audible, even with shunt regulation with CCSs (I am using a similar arrangement with series 0A2s for a 300V B+). I would recommend using as small value as you can get away with without comprimising the hum or hash on the outputs). If using electrolytic caps, bypass them with signal quality film or oil caps.
That being said, the Aikido circuit is capable of very low level resolution, the naturally low PSRR means that meticulous attention to PS quality will reap great benefits in resolution.
Just one comment to Dave Dove, regarding the problem I was having a few weeks ago concerning excessive noise that I thought was attributed to the safety resistor divider, I subsequently found that the 12FQ7s I was using at the time had a problem with poor heater insulation that was producing excessive buzz noticable through my 105dB efficient horn speakers. Using a different tube solved the problem. Thanks for all the thought and effort in your comments to me on the problem.
David
Dave, nice thanks for the reply 6N1P are pin compatible i take it? I might give them a try, what cathode resistors do you use with them and HT voltage.
dmcgown,
I will certainly add a CCS then, definatly solid state because i dont have much room left inside my case after all those regulators.
My PSUcaps (~80uf per channel) consist of 3 smaller caps in // with a 340nf cap accross their output.
I think Jfets are a good solution to the thump, more compact and reliable than relays but to get the output muted on turnoff might be tricky.
Thanks for your posts.
Craig
dmcgown,
I will certainly add a CCS then, definatly solid state because i dont have much room left inside my case after all those regulators
.My PSUcaps (~80uf per channel) consist of 3 smaller caps in // with a 340nf cap accross their output.
I think Jfets are a good solution to the thump, more compact and reliable than relays but to get the output muted on turnoff might be tricky.
Thanks for your posts.
Craig
dmcgown:
"That being said, the Aikido circuit is capable of very low level resolution, the naturally low PSRR means that meticulous attention to PS quality will reap great benefits in resolution."
I don't understand your comment. I thought the whole point of the Aikido concept was to have a very good PSRR.
"That being said, the Aikido circuit is capable of very low level resolution, the naturally low PSRR means that meticulous attention to PS quality will reap great benefits in resolution."
I don't understand your comment. I thought the whole point of the Aikido concept was to have a very good PSRR.
I have been working on a SN7 Aikido breadboard.
Very, very good sound.Open and airy sound. Very transperent.
I used another of Broskie's ideas as well, works great. 18.9 volt heater transformer with 21 ohm resistors on each valve. I can plug in any 6,8 or 12 volt SN7 or such. Then I used 19AX7 damper diodes off the same transformer. These diodes are cheep as chips and sound good.
I found the power supply quite critical. The sound of caps quite noticeable. I now have ..10H , 10uf Solen, 10H, 20uf Solen, 500ohms 30uf Solen, 10H, 50uf motor run, 10H 100uf motor run.Sounded better than NOS oils or Cerafines and or a less complex supply.
Mucking around with volume pots was interesting. Vishay 1M cermet law faked with carbon comp 50k is very transperant. Will try it as a series conection next.
Cheers.
Very, very good sound.Open and airy sound. Very transperent.
I used another of Broskie's ideas as well, works great. 18.9 volt heater transformer with 21 ohm resistors on each valve. I can plug in any 6,8 or 12 volt SN7 or such. Then I used 19AX7 damper diodes off the same transformer. These diodes are cheep as chips and sound good.
I found the power supply quite critical. The sound of caps quite noticeable. I now have ..10H , 10uf Solen, 10H, 20uf Solen, 500ohms 30uf Solen, 10H, 50uf motor run, 10H 100uf motor run.Sounded better than NOS oils or Cerafines and or a less complex supply.
Mucking around with volume pots was interesting. Vishay 1M cermet law faked with carbon comp 50k is very transperant. Will try it as a series conection next.
Cheers.
Sorry about the confusion in my statement about PSRR, Bas correctly noted my intention, the Aikido having high PSRR. My intent was to note that good attention to power supply will allow you to get even better low level resolution. Obviously, it is subject to the law of diminishing returns, but this hobby is about excess anyway.
As far as a solid state CCS, I can recommend the DN2540 cascode current source that Bas has on his website. Low parts count, easy to put together, and sounds good! One little trick I found, use the TO-220 version for the top and the TO-92 version on the bottom, they line up pin for pin with either jumpers or resistors between the leads. I have a little three pad wide layout that I can cut off a DIP style prototyping PC board like what is available at Rat Shack. It is 1 cm wide by 3 cm long. I just attach the upper MOSFET to my chassis or top plate using mica insulator (with suitable insulated bolt), and it supports the entire CCS nicely.
David
As far as a solid state CCS, I can recommend the DN2540 cascode current source that Bas has on his website. Low parts count, easy to put together, and sounds good! One little trick I found, use the TO-220 version for the top and the TO-92 version on the bottom, they line up pin for pin with either jumpers or resistors between the leads. I have a little three pad wide layout that I can cut off a DIP style prototyping PC board like what is available at Rat Shack. It is 1 cm wide by 3 cm long. I just attach the upper MOSFET to my chassis or top plate using mica insulator (with suitable insulated bolt), and it supports the entire CCS nicely.
David
The most important design aspect of the Aikido is to incorporate PSU noise into the signal for cancellation so it provides a very high PSSR. This should reduce the demand on the quality of the PSU. A high PSSR means the amp is less subject to the variations of impedance of the PSU at different frequencies and less subject to PSU ripples.
I have been very interested in the Aikido as I believe that even using a very cheap, simple PSU consisting of toroidal Tx, SS diodes, CRC filters the Aikido should still perform very well and there should be little advantage of using very expensive parts on the PSU or using a fancy regulator such as CCS fed shunt regulator.
Since someone recommends using high quality PSU for the Aikido, it would be good for the guys who have built it to verify the theory by reducing the current PSU to make a comparison of sound and report the results back to the forum.
I am a firm believer of PSU in conventional topologies. I have built a simple 12B4A line preamp and it sounds bad when using bad caps and sounds absolutely wonderful when using quality caps on the PSU. However, I don't think a high quality PSU is necessary for the Aikido simply because it is an Aikido. For that reason I want to build it and I have some cheap redundant parts I can use.
Regards,
Bill
I have been very interested in the Aikido as I believe that even using a very cheap, simple PSU consisting of toroidal Tx, SS diodes, CRC filters the Aikido should still perform very well and there should be little advantage of using very expensive parts on the PSU or using a fancy regulator such as CCS fed shunt regulator.
Since someone recommends using high quality PSU for the Aikido, it would be good for the guys who have built it to verify the theory by reducing the current PSU to make a comparison of sound and report the results back to the forum.
I am a firm believer of PSU in conventional topologies. I have built a simple 12B4A line preamp and it sounds bad when using bad caps and sounds absolutely wonderful when using quality caps on the PSU. However, I don't think a high quality PSU is necessary for the Aikido simply because it is an Aikido. For that reason I want to build it and I have some cheap redundant parts I can use.
Regards,
Bill
Hi Bill,
IMHO I think that you are totally wrong about the power supply.
PSRR is only one thing that the Aikido is better at. Ripple is a small part of a power supplies sound.It is very easy to hear a better power supply with the Aikido circuit. I have tried many variations.The "better" supply always sounded better.The difference is not supple.
IMHO I think that you are totally wrong about the power supply.
PSRR is only one thing that the Aikido is better at. Ripple is a small part of a power supplies sound.It is very easy to hear a better power supply with the Aikido circuit. I have tried many variations.The "better" supply always sounded better.The difference is not supple.
i'm a newby, have only been diy for a couple of years. find it fascinating and addicting as well as bewildering at times. it's one thing to read and build from a schematic, but quite another to know why things work the way they do. hopefully someone here can help me with my aikido!
mine uses 12au7s into 12bh7s(this combo simply because it's what i had laying around, i was determined to spend no money on this!) i used JBs schematic from his first artical with the saftey dividers. B+ is @263, the divided voltages are all 129v. 12au7 bias at 3.5 v for about around 7 ma. 12bh7s bias at 3.5 and one just under 4 for about 12-15ma.(i havnt directly measured current yet, i roughed current out from data)
the problem, WAY to much gain. i used a 100k fako alps stepped pot instead of 50k. i changed nothing else, followed the schema exactly. could the pot be the problem? in a fast measurement of ac at output rcas with vol. turned up i saw peaks of over 5v! (used decent b&k dmm to check)plays quite loudly at only first click on vol.
anyone have any thoughts ? if not the pot are there other ways to reduce gain without changing tubes? sound seemed pretty decent for the little time i listened, but its to loud!!
mine uses 12au7s into 12bh7s(this combo simply because it's what i had laying around, i was determined to spend no money on this!) i used JBs schematic from his first artical with the saftey dividers. B+ is @263, the divided voltages are all 129v. 12au7 bias at 3.5 v for about around 7 ma. 12bh7s bias at 3.5 and one just under 4 for about 12-15ma.(i havnt directly measured current yet, i roughed current out from data)
the problem, WAY to much gain. i used a 100k fako alps stepped pot instead of 50k. i changed nothing else, followed the schema exactly. could the pot be the problem? in a fast measurement of ac at output rcas with vol. turned up i saw peaks of over 5v! (used decent b&k dmm to check)plays quite loudly at only first click on vol.
anyone have any thoughts ? if not the pot are there other ways to reduce gain without changing tubes? sound seemed pretty decent for the little time i listened, but its to loud!!
gmcgown,
Thanks for the tip but those transistors are very hard for me to get around here. (bas actually offered me some of his in another thread about CCS's). Ill probably grab whatever is available to me and stick it in cascode, probably end up being bipolar or using IRF820's as they are common for me.
Peter Clements,
Thanks for sharing your results, that PSU sounds good!
Bill,
I think it is important to get a good supply behind this circuit even though it cancels PSU noise in theory.
It is impossible to perfectly match triodes in each stage as this circuit requires and cancel all the noise.
So why not make it easier for the circuit by giving it as little noise as possible and have it reject the same ratio of that.
Cheers
Craig
Thanks for the tip but those transistors are very hard for me to get around here. (bas actually offered me some of his in another thread about CCS's). Ill probably grab whatever is available to me and stick it in cascode, probably end up being bipolar or using IRF820's as they are common for me.
Peter Clements,
Thanks for sharing your results, that PSU sounds good!
Bill,
I think it is important to get a good supply behind this circuit even though it cancels PSU noise in theory.
It is impossible to perfectly match triodes in each stage as this circuit requires and cancel all the noise.
So why not make it easier for the circuit by giving it as little noise as possible and have it reject the same ratio of that.
Cheers
Craig
bas-thanks!, thats all it was! i whipped the thing together yesterday and in my rush to hear it, screwed up.
wow this thing sounds really great! better than the audio experience 12ax7 line i've been useing. lots of detail, wieght and good deep soundstage,very liquid. it is a little noisy, but my ps is pretty basic, 5ar4 to 7h-50u-7h-50u. it was borrowed from a el84 amp project thats on hold. i'm also running ac heaters that aren't referenced to any voltage. will changing to dc be worth the work? is referencing them to a voltage rather than ground advisable?
thanks! charlie
wow this thing sounds really great! better than the audio experience 12ax7 line i've been useing. lots of detail, wieght and good deep soundstage,very liquid. it is a little noisy, but my ps is pretty basic, 5ar4 to 7h-50u-7h-50u. it was borrowed from a el84 amp project thats on hold. i'm also running ac heaters that aren't referenced to any voltage. will changing to dc be worth the work? is referencing them to a voltage rather than ground advisable?
thanks! charlie
Your psu is almost state of the art With7h-50uF-7h-50uF there should not be a lot of noise.
You can always try dc ofcourse..but first reference the ac to earth...for now. Create a centre tap with 2 50R resistors...and from the centre tap to your earth.. I prefer star earth. This should make your preamp a lot quieter.
Then create a reference voltage for the heater/cathode insulation spec (which should not be "violated" for too long...takes months apparently for damage to occur. )
b+ -> 200k -> (your reference voltage) cap to earth -> 100k to earth..
Now connect the psuedo centre tap you made earlier to this voltage reference.
Good luck..and be careful.
Regards,
Bas
With7h-50uF-7h-50uF there should not be a lot of noise.You can always try dc ofcourse..but first reference the ac to earth...for now. Create a centre tap with 2 50R resistors...and from the centre tap to your earth.. I prefer star earth. This should make your preamp a lot quieter.
Then create a reference voltage for the heater/cathode insulation spec (which should not be "violated" for too long
...takes months apparently for damage to occur. )b+ -> 200k -> (your reference voltage) cap to earth -> 100k to earth..
Now connect the psuedo centre tap you made earlier to this voltage reference.
Good luck..and be careful.
Regards,
Bas
Sorry if some feel this is threadjacking, but are you guys serious about these "noise cancels in theory" comments? I'm not an EE, but from what I know, this sounds like hogwash. I would think that even a simple tube circuit is not purely deterministic and so neither noise nor any other signal component could not be manipulated in such a manner. Feel free to call me stupid if this is just over my head.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.