I know that this has been discussed many times before, but here is a dilemma:
My Aikido works, albeit a small degree of hum. This caused me to go back and investigate whether my PSU is working correctly. The current set-up is this:
1. 125-0-125 transformer, giving 250 end-to-end
2. bridge rectifier with 4 X 1N4007
3. 1.5uF cap
4. 10H 102ohm choke
5. 60uF cap
5. 10H 102 ohm choke
6. 60uF cap.
Now, if I simulate this in PSUDII, which I did before I purchased the ironware, Iget around 296VDC with just under 1mV ripple.
So, why does it not work this way with my Aikido?? Maybe, I am not drawing quite enough milliamps for a 10H choke, and that the effect of this is not modelled in PSUDII. My 6sn7s pull a total current of 20 - 24 mA.
If I model the same PSU, but replace the 10H with 20H 181 ohm, I also get about 300VDC, but with only 0.2mV ripple.
So, should I replace my 10H's with 20H's? Given that minimum current requirements for a choke-input filter is VDC / L, then 300 / 20 = 15......ie. 15mA minimum current draw. Seeing as my 6sn7's draw about 24mA, I am pulling at least enough for the choke to work properly.
My gut feeling is that if I spend a around $60 on a pair of new chokes, safe in the knowledge that my tubes will draw enough current to make them work properly, without having to use a bleeder resistor to increase the current draw, I'll be able to sleep more soundly, and hopefully with an amp having zero hum issues.
Please advice....
Charlie
My Aikido works, albeit a small degree of hum. This caused me to go back and investigate whether my PSU is working correctly. The current set-up is this:
1. 125-0-125 transformer, giving 250 end-to-end
2. bridge rectifier with 4 X 1N4007
3. 1.5uF cap
4. 10H 102ohm choke
5. 60uF cap
5. 10H 102 ohm choke
6. 60uF cap.
Now, if I simulate this in PSUDII, which I did before I purchased the ironware, Iget around 296VDC with just under 1mV ripple.
So, why does it not work this way with my Aikido?? Maybe, I am not drawing quite enough milliamps for a 10H choke, and that the effect of this is not modelled in PSUDII. My 6sn7s pull a total current of 20 - 24 mA.
If I model the same PSU, but replace the 10H with 20H 181 ohm, I also get about 300VDC, but with only 0.2mV ripple.
So, should I replace my 10H's with 20H's? Given that minimum current requirements for a choke-input filter is VDC / L, then 300 / 20 = 15......ie. 15mA minimum current draw. Seeing as my 6sn7's draw about 24mA, I am pulling at least enough for the choke to work properly.
My gut feeling is that if I spend a around $60 on a pair of new chokes, safe in the knowledge that my tubes will draw enough current to make them work properly, without having to use a bleeder resistor to increase the current draw, I'll be able to sleep more soundly, and hopefully with an amp having zero hum issues.
Please advice....
Charlie
Hi Charlie,
I think your problem lies elsewhere. The power supply you show should be way more than adequate. Throwing more Henrys at it will not help. I am running an Aikido with a real POS power supply, 0 Henrys and a $6 transformer. Look at it with a scope and say NO way! It is dead silent and sounds great. The Aikido has exceptional power supply noise rejection.
Your noise is either some sort of ground loop or from the input. Have you disconnected everything and shorted all of the inputs? this is important, just unplugging will not do. Get a set of Radio Shack plugs and solder the two pins together (Center to shield/ground) With these in place. How is the hum? Dose it vary with the volume control? Is it gone when the vol is all the way up? If you pull the AC plug dose it stop right away or fade out?
Do this in order and report back…John
I think your problem lies elsewhere. The power supply you show should be way more than adequate. Throwing more Henrys at it will not help. I am running an Aikido with a real POS power supply, 0 Henrys and a $6 transformer. Look at it with a scope and say NO way! It is dead silent and sounds great. The Aikido has exceptional power supply noise rejection.
Your noise is either some sort of ground loop or from the input. Have you disconnected everything and shorted all of the inputs? this is important, just unplugging will not do. Get a set of Radio Shack plugs and solder the two pins together (Center to shield/ground) With these in place. How is the hum? Dose it vary with the volume control? Is it gone when the vol is all the way up? If you pull the AC plug dose it stop right away or fade out?
Do this in order and report back…John
John,
I hear you. Since my multimeter has died, and my PSU is unsoldered, I'll do as you recommend tomorrow. I have already done a number of different combinations to lower the noise, in fact when I took the ground from the PSU and ran it to the PCB GND, then took the combined PCB / PSU supply to a single point on the chassis, things really improved. It got even better, when I ran this ground through a 10 ohm resistor, paralleled with cap, paralleled with diode. This lifted the ground slightly, the cap adding some RF filtering, and the diode bridge acting as extra safe route to GND should the resistor and cap fail.
I guess is that I could increase the resistor value slightly, say to 20 or even 30 ohms.
The problem is that I have done so many things to the PSU, that I cannot remember what the voltages were at various stages in its production. At first, I was getting around 425V, which was ridiculous, then I got it to drop to more like 370V. A few days ago, it seemed like I was only getting 200V, but that could be my dying meter.
As modelled, I should get around 300V, although when reading through Morgan Jones, he does say that if not enough current is drawn, a choke input filter will yield up towards V * sqrt2, because the choke is not functioning properly. I am using a small value cap at the input, but my filter may now be behaving somewhere between cap and choke input.
Charlie
I hear you. Since my multimeter has died, and my PSU is unsoldered, I'll do as you recommend tomorrow. I have already done a number of different combinations to lower the noise, in fact when I took the ground from the PSU and ran it to the PCB GND, then took the combined PCB / PSU supply to a single point on the chassis, things really improved. It got even better, when I ran this ground through a 10 ohm resistor, paralleled with cap, paralleled with diode. This lifted the ground slightly, the cap adding some RF filtering, and the diode bridge acting as extra safe route to GND should the resistor and cap fail.
I guess is that I could increase the resistor value slightly, say to 20 or even 30 ohms.
The problem is that I have done so many things to the PSU, that I cannot remember what the voltages were at various stages in its production. At first, I was getting around 425V, which was ridiculous, then I got it to drop to more like 370V. A few days ago, it seemed like I was only getting 200V, but that could be my dying meter.
As modelled, I should get around 300V, although when reading through Morgan Jones, he does say that if not enough current is drawn, a choke input filter will yield up towards V * sqrt2, because the choke is not functioning properly. I am using a small value cap at the input, but my filter may now be behaving somewhere between cap and choke input.
Charlie
O.K. what this is doing is somewhat lifting a ground loop.
Time for some rip and tear. First who’s board are you using and are you willing to start fresh?
If so go buy a new meter Now is the time to pull them out of the case and play with them. You need to have a level of confidence in your equipment and technique. There is nothing worse then wondering is this right or am I doing it wrong. A few reality checks are handy. Things to watch for are, looking for 300V OK if it’s there but if it’s not. you might see 200-300 MILLA VOLT this has got me, keep a 9 V battery close by for a sanity check and your filaments if AC will be in real volts AC .
For resistance get a few known resistors. Keep away from the 10R,100R.1k. This is where meter decades changes Get say a 15R or 47R, a 150R or 470R . Try to get one Just under 1k and just over 1 k This is were DVM ‘s can lead you astray but if you have a 990R and a 1100R (1.1k ) or so well you know.
Next go and buy a copy book, I like the ones with graph paper, they make me feel more like a scientist
John,
The board that I am using is Broskie's setero octal board. I have checked and rechecked values of components. Also, I did the sensible thing of checking resistor values prior to stuffing. I also used Vishay-Dale resistors throughout as they have a pretty good tolerance. My resistor choices were based on Broskie's recommendations for 6sn7s. I even went back and resoldered all parts, just in case I had cold solder.
I believe that my problems are either with the PSU or the grounding, and not the PCB. In fact, two things make me think that both are contributing to the hum:
1. When I changed location of the grounding to utilize the star-earth of the PCB, the level of hum diminished significantly.
2. When I added the 10 ohm resistor onto that ground (between star-earth and chassis), hum dropped another notch. This reduced the hum to a level that was only discernable during quiet pieces of music.
3. The PSU may also be adding some hum, if I am drawing only 24mA, or so, and a 10H choke really does require 30mA to begin working properly, then it could be filtering inadequately. However, I vaguely remember reading that PSU problems tended to result in 120Hz hum, while I certainly seem to only have 60Hz hum. I haven't measured this, but have used a software synth to create these frequencies (and with a few different waveforms).
I will invest in a new meter this morning and re-wire my PSU and check it. I am very reluctant to do the "plug-pull" test, as I am using a solidstate power amp and don't want to risk any damage.
My re-wired PSU uses soldered connections throughout, while I had used spade connectors before - this may cause some improvement. If I still get some hum, I'll add two 20K 10W resistors in series across the PSU output to draw a little more current (around 10mA, I think). Then, if hum is still present, I'll replace the 10 ohm ground-lift resistor with 20 ohm.
I am feeling more confident this morning.
Thanks for the advice.
Charlie
The board that I am using is Broskie's setero octal board. I have checked and rechecked values of components. Also, I did the sensible thing of checking resistor values prior to stuffing. I also used Vishay-Dale resistors throughout as they have a pretty good tolerance. My resistor choices were based on Broskie's recommendations for 6sn7s. I even went back and resoldered all parts, just in case I had cold solder.
I believe that my problems are either with the PSU or the grounding, and not the PCB. In fact, two things make me think that both are contributing to the hum:
1. When I changed location of the grounding to utilize the star-earth of the PCB, the level of hum diminished significantly.
2. When I added the 10 ohm resistor onto that ground (between star-earth and chassis), hum dropped another notch. This reduced the hum to a level that was only discernable during quiet pieces of music.
3. The PSU may also be adding some hum, if I am drawing only 24mA, or so, and a 10H choke really does require 30mA to begin working properly, then it could be filtering inadequately. However, I vaguely remember reading that PSU problems tended to result in 120Hz hum, while I certainly seem to only have 60Hz hum. I haven't measured this, but have used a software synth to create these frequencies (and with a few different waveforms).
I will invest in a new meter this morning and re-wire my PSU and check it. I am very reluctant to do the "plug-pull" test, as I am using a solidstate power amp and don't want to risk any damage.
My re-wired PSU uses soldered connections throughout, while I had used spade connectors before - this may cause some improvement. If I still get some hum, I'll add two 20K 10W resistors in series across the PSU output to draw a little more current (around 10mA, I think). Then, if hum is still present, I'll replace the 10 ohm ground-lift resistor with 20 ohm.
I am feeling more confident this morning.
Thanks for the advice.
Charlie
Charlie,
There is an Aikido thead that Bas had on a group buy I believe. Anyway he drew out the transformer and wiring in order to make it easy for people who hadn't constructed something before. I realize you have done a number of projects in the past but it might serve as an example. Maybe you are overlooking the obvious and aren't seeing the problem. As mentioned above I don't think you need a larger choke to reduce the hum.
There is an Aikido thead that Bas had on a group buy I believe. Anyway he drew out the transformer and wiring in order to make it easy for people who hadn't constructed something before. I realize you have done a number of projects in the past but it might serve as an example. Maybe you are overlooking the obvious and aren't seeing the problem. As mentioned above I don't think you need a larger choke to reduce the hum.
I agree it’s probably not the board.
While you in the power supply. Lets try a few things. First remove the chokes and replace them with resistors 1k 2w or so. The second choke in particular is not doing much.
This will lower your voltage but should be enough for testing. You can restore the voltage by removing the first cap and resistor just go 60uf 1k 60uf if you have a larger cap on hand for the second position it wouldn’t hurt
While we’re at it add a .01 600v or better cap across the output of your trans before the diodes, this is to shunt there noise not part of the current issue.
Next how do your filaments? If AC I would recommend putting 300k & 100k resistor in series and connecting them across the output of B+ with the 300k to + and the 100k to ground now connect your heater winding center tap to the middle. If your winding doesn’t have a CT make one. Connect a 100R resistor to each leg of the winding, connect the other ends together this becomes your CT. This will float your filaments at about 100V or so and will keep the top half of the tubes within there heater/cathode limit. It also acts as a bleeder for your PS…John
While you in the power supply. Lets try a few things. First remove the chokes and replace them with resistors 1k 2w or so. The second choke in particular is not doing much.
This will lower your voltage but should be enough for testing. You can restore the voltage by removing the first cap and resistor just go 60uf 1k 60uf if you have a larger cap on hand for the second position it wouldn’t hurt
While we’re at it add a .01 600v or better cap across the output of your trans before the diodes, this is to shunt there noise not part of the current issue.
Next how do your filaments? If AC I would recommend putting 300k & 100k resistor in series and connecting them across the output of B+ with the 300k to + and the 100k to ground now connect your heater winding center tap to the middle. If your winding doesn’t have a CT make one. Connect a 100R resistor to each leg of the winding, connect the other ends together this becomes your CT. This will float your filaments at about 100V or so and will keep the top half of the tubes within there heater/cathode limit. It also acts as a bleeder for your PS…John
John,
It seems like I got things going again. One problem I noticed is that out AC outlets are giving 123VAC, which is considerably higher than I had anticipated. This means that the 125-0-125 transformer actually puts out more like 295VAC end-to-end.
With the 1.5uF cap, I was getting just around 420VDC, which is more than PSUDII suggested.
Anyway, I took out the first cap and went choke input. I get 260VDC, which is pretty much what the 6sn7's are designed for. Anyway, I still got the same level of hum, works with loud music, is not volume dependant, etc.
I replaced the 10 ohm ground lift with a 20 ohm and now I actually have to put my ear to the speaker in order to hear a faint hum. I'll get a 30 ohm 5W and see if that totally diminishes it, although I am not going to rush out and do that as I need to redo the signal wiring first.
I am going to post this message as a new topic, as I also want to ask how I can find out if my chokes are designed for choke input duty.
Regards,
Charlie
It seems like I got things going again. One problem I noticed is that out AC outlets are giving 123VAC, which is considerably higher than I had anticipated. This means that the 125-0-125 transformer actually puts out more like 295VAC end-to-end.
With the 1.5uF cap, I was getting just around 420VDC, which is more than PSUDII suggested.
Anyway, I took out the first cap and went choke input. I get 260VDC, which is pretty much what the 6sn7's are designed for. Anyway, I still got the same level of hum, works with loud music, is not volume dependant, etc.
I replaced the 10 ohm ground lift with a 20 ohm and now I actually have to put my ear to the speaker in order to hear a faint hum. I'll get a 30 ohm 5W and see if that totally diminishes it, although I am not going to rush out and do that as I need to redo the signal wiring first.
I am going to post this message as a new topic, as I also want to ask how I can find out if my chokes are designed for choke input duty.
Regards,
Charlie
cbutterworth said:
My experience was that the main contribution was not from the PSU (which should be the strong point of the aikido approach anyway).
http://www.diyaudio.com/forums/showthread.php?postid=992945#post992945
I had actually hooked up the aikido preamp to an external regulated DC power supply (10mV out of 300V) and the hum remained about the same. I was only able to reduce it a little bit by wiring/grounding rearrangement (trial and error) but still below my expectation. I accepted that the small hum probably would be very difficult to get rid of, and learning to live with it. My aikido definitely is not dead silent (other members do have much better luck and/or skill I suppose).
cbutterworth said:
5W is serious overkill for the lift resistor. There should be no current through that resistor. If you're are worried about it failing open in a fault condition, parallel it with a pair of antiparallel diodes (diodes with enough current capacity to blow your fuse before they do).
I built a RIAA phono preamp using an Aikido topology on the output. It has no hum. So any audible hum from your lower gain preamp is something that can definitely be fixed. On thing you might check is the resistor divider the supplies the lower grid on the output tube (the noise cancelling input). You could plug a pot in here and adjust for lowest noise to see if that improves matters. This is probably not the cause, but it's easy to check.
Sheldon
Charlie
As others have pointed out (meaning me) Your current power supply is good but not best. I told you in great detail how to make it much closer to the very best for cheap. Your lack of faith is painful : )
My Aikido phono is dead quiet at full gain. With your identical supply my line amp was dead quiet. Your looking for love in all the wrong places! Neither increasing the first cap nor the choke(s) is going to get you where you ultimately want to go. OK, I'm done now. Best of luck whichever way you go.
As others have pointed out (meaning me) Your current power supply is good but not best. I told you in great detail how to make it much closer to the very best for cheap. Your lack of faith is painful : )
My Aikido phono is dead quiet at full gain. With your identical supply my line amp was dead quiet. Your looking for love in all the wrong places! Neither increasing the first cap nor the choke(s) is going to get you where you ultimately want to go. OK, I'm done now. Best of luck whichever way you go.
JoshK said:
In case someone else (like me) is trying to build a PS for the Aikido PCB, here is a link to the other post refered to above.
Related Post from Bruce
Ron
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.