• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Closed loop stability help needed.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Karpov stated it was good in triode mode is why I tried it. I'm intrigued by the shift in high order and especially odd order harmonics with bias current. I've never read that this happens. It is possible to adjust for 3rd and 5th high with 7th and 9th low, or compromise and lower all compared to the corresponding higher even harmonic.

I'm disappointed in the Mu Stage distortion. I expected it to be much lower, especially for the second harmonic. I may try a drive stage with a simple current source anode load and LED in the cathode to see how it drives the 6P41S for comparison. It will have to be cap coupled though, and I was trying to avoid that.
 
I believe I may have figured out what is happening.

I think this may be an extreme case of harmonic cancellation, wherein under certain bias conditions the harmonics produced by the Mu stage are interacting with the harmonics generated by the 6P41S resulting in an extreme cancellation effect.

Changing the bias of the 6P41S is changing the harmonics it produces, and as I adjust it's bias, I can see the resulting cancellation.

How would I test to prove or disprove this?

If I change the bias of the MuStage, I can eliminate the effect. This strongly suggests I am correct, but does not prove it.
 
Last edited:
I removed the servo, and reconfigured the amp as a triode strapped 6J9P with red LED in cathode and current source anode load as the input stage.

The 6P41S was switched to fixed bias.

The problem at 20KHz is certainly external as the first image shows the amp with power off and the noise is still there.

Second image is amp with UL output, at 1W.

I'll work on triode mode this weekend if I can get time.
 

Attachments

  • TestSet_W_Power_Off.JPG
    TestSet_W_Power_Off.JPG
    86.1 KB · Views: 167
  • TriodeStrapped_6J9P_Input.JPG
    TriodeStrapped_6J9P_Input.JPG
    91.1 KB · Views: 174
despur the input tube by placing 20-220pf cap between grid and anode of the input tube. also shunting the led with a 220pf cap helps too.



limit rf amplification in the servo circuit by adding a 220 pf cap across the collector-emmiter pins.

if you do all of this you shouldn't have oscillation issues. also lowering the input grid bias from 1M to say 860-470K helps in some tubes when you put a big input cap on them.
 
Last edited:
The computer is an old desktop, and I have no idea what the switching frequency of the power supply is. I could take a loop antenna to it and try to find out.

I had hoped to try a USB sound card with my laptop, but from your comments it does not sound promising.

I'll try DNHs suggestions back to the original circuit when I can get back on the workbench (helping son move and tied up with that for a while).
 
I finally tracked down the bump from 10-20KHz. It is caused by florescent lights in the den. They are "Daylight" 5500K lamps for SAD so I can't get rid of them.

I guess the LED lamps are going to be almost as bad as florescent bulbs since they use switching regulators to run them as well. The gas discharge in the florescent lamps probably contributes as well so they are worse in that regard and LED lamps won't have that problem (not to mention no mercury).

The upward swing from 1KHz to 20KHz is most likely additional switching devices dumping noise on the power line which will be very difficult to filter out.

I tried to install the EMU0404 on the laptop and finally got it to work, but can't get AudioTester to play nicely. It can't find the registration, and I can't find the directory where it belongs. So, I may install the EMU-0404 on the shop computer to see how it compares to the M-Audio 192.

More to come on amp testing later.
 

Attachments

  • W-Florescent.JPG
    W-Florescent.JPG
    103.2 KB · Views: 137
  • W-Florescent-off.JPG
    W-Florescent-off.JPG
    113.5 KB · Views: 142
Last edited:
I'm intrigued by the shift in high order and especially odd order harmonics with bias current......I think this may be an extreme case of harmonic cancellation

I noticed a similar effect several years back with a simple triode driving triode SE amp. The odd order harmonics were affected by the bias current and at a particular current the third could be reduced by 30db or more. I rolled tubes, caps, and just about every part on the bench before finding out that the critical component in all of this was the OPT. The little Edcor XSE could nearly zero H3 at 80 mA regardless of the tube it was connected to. Other OPT's had this effect at different currents but none were as pronounced as the Edcor.

It would be interesting to see if the effect was constant across frequency, or changed. I am guessing this is due to the change in OPT inductance with bias current.
 
I'm using Edcor transformers, GXSE10-8-2.5K, GXSE15-8-3.5K, XSE10-8-4K, and GXSE10-8-5K.

The GXSE10-8-5K gives the lowest overall distortion.

The GXSE15-8-3.5K giving the lowest distortion at 20Hz due to it's larger core.

I'll play with bias current and change test frequency for the FFT and see what shows up.
 
float you heater supply, take a .1uf cap and go from signal ground to the + of your dc supply. I think the heater is throwing trash into the tube elements. I haven't found a documented name for it but I call this electromagnetic impression interference. To test your tube circuit remove all voltages except the heater. take an oscilloscope and look at the cathodes of all the tubes. Ideally, there should be nothing but a strait line on your scope. If not do the cap trick. This does wonders for finding the hum in a damage/ill configured heater circuit.
 
I used a resistor divider from B+ to ground to set a 60V reference for biasing the heaters. I also set up a separate divider for the bottom of the Mu stage when I was testing it, and biased it to be 60V above it's cathode (on a different heater transformer).

However, I did not bypass them with caps, which appearantly does help a great deal.

A 1uF in parallel with a 0.033uF cap reduces it considerably.
 

Attachments

  • W-bypass_cap.JPG
    W-bypass_cap.JPG
    85.5 KB · Views: 121
what are you using for heaters power supply? ac with a transformer or a dc supply out of a computer tower? if your trying to use a laptop power supply, you have to filter that pulsed dc that comes out of those types of power supplies.
you can troubleshoot if it is the heater circuit by substituting it with a motorcycle battery with its neg lead tied to the B+ reference.
 
Last edited:
I tried 220, 440, 1K, 2K,3K,and 4K as I could see enough harmonics to be useful.

In all cases above 440Hz, the fft looked identical at 65mA with all harmonics shifted. If I decreased or increase bias current , third harmonic decreases at the expense of 5th harmonic for a bit, then turned up and both third and 5th increase along with all other harmonics.

at 220 and 440Hz, third is very dependent on bias current, but 5th is not effected near as much. As bias current decreases, third harmonic decreases.

So, I am guessing that at low frequencies (below 1KHz) harmonic distortion is more of a function of magnetic flux, and at frequencies (1KHz and above) it becomes more dependent on coupling capacitance/Leakage inductance.
 
DNH, I am not using the computer to provide power to anything. It is there to run the audiotester program for FFT analysis.

The amp is powered by two transformers for heaters, and one for B+/B-. The power is CLC filtered and then regulated to +/- 250V.

Was "Dave's not here" on the Big Bamboo album? Or "Up in smoke"? Cheech and Chong were great.
 
ok, I can picture the heater circuit now that you should have. two 1W 120 ohm resistors across the heater winding. this will make a virtual ground point for your heater winding. at that center tap, a wire is placed to the voltage divider network. two .1 uf caps, one side of the cap is tied on each end of the heater windings, the other side star ground. the B+ and B- has a .1 uf cap tied to them and the star ground. now any noise that is happening on your powers should be gone. also, are you using a capacitive probe? I suspect a possible ground loop with your diagnostic equipment but since the unit is malfunctioning, I'm not going that direction. A .1 cap is a good cap to use in between signal injection diagnostic devices and .01 in rf circuits. O-scopes can couple AC by selecting a switch.
I got that nick name in the music world, along with: "Little Davey","Davey", "Drtechno", "Dr.T", and "Mr. Wizzard "
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.