• 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.

SE OPT air gap vs saturation

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello everyone.
Is it possible to get a higher saturation frequency when increasing the air gap? This doesn't make any sense to me, but it seems to be happening.

I'm testing a new little SE OPT I've wound, for a low power amplifier.
Sounds very good with an air gap of .15 mm, I found by numerical methods.

Now I'm trying to adjust the air gap, taking P.Turner's method as inspiration.
Basically:


- With a dummy load, I find the maximum output voltage before clipping
- Then remove the load from the secondary.
- Re-adjust the output level.
- Decrease input signal frequency, looking at output voltage and anode current in the scope, until saturation of the core is shown.

I got Fsat = 38 Hz and Fc = 16 Hz.

So I wanted to try to increase the gap, to get lower Fsat, and higher Fc.
But, while Fc does climb, Fsat does too,...


Attached are two of the test I've done: 0.15 mm and 0.4 mm. Blue trace is output voltage. You can see a small hump, I assume, due to core saturation. If I go down in frequency, this hump gets much more prominent, until totally deforming the sinewave.

How is it possible that saturation occurs at higher frequencies while increasing the gap? :confused:

I tried four different air gaps, and all results relate.
 

Attachments

  • se_opt_fsat.png
    se_opt_fsat.png
    27.4 KB · Views: 388
Last edited:
The tubes are probably unable to deliver enough current in such a low inductance.

It makes a lot of sense,... the ac current on both cases is very close to the dc bias current when the "hump" appears on the output voltage. Maybe the output tube is near cutoff.

If this is the case, what I thought was a saturation frequency of 33 Hz for the smaller gap, is not really saturation. Which, again, seems to be the more reasonable explanation.

Even though I have no much room left until reach max anode dissipation, I will try to bias a little hotter, and come back with some results.
 
. . .
Even though I have no much room left until reach max anode dissipation, I will try to bias a little hotter, and come back with some results.
This will increase the DC induction leaving less room for AC (the signal) :( Remember that DC part of induction is proportional to DC current multiplied by the number of turns and decreases with the gap while the AC part decreases with the number of turns and the frequency but as nothing to do with the gap. The sum of the two parts must remain below the value where the iron begins to saturate (in fact, looses its permeability and killing primary inductance).

The door is narrow !
 
At low frequency the main problem is the L that collapse so the distortion is mainly from the tube that can't deliver current due the lower Z. The problem of the dc current, in a good transformer, is not so important as the L that it is (due the core material and topology of winding) effectively measured.

In the other hand the parasitic at high frequency are more important. If we can solve the problem, less or more, at low frequency with the dimensions of the nuclues ( one the most important parameter, of course) for the high frequency a good architecture and a manual capabilities on winding are more heavy and diffcult to find.

But in some case the door is not so narrow! :)

This is my opinion
Walter
 
Last edited:
This will increase the DC induction leaving less room for AC (the signal) :(

Yes Yvesm. Bias hotter was just testing, to see if I could prevent the output tube to go into cutoff.

The answer is no. The tube runs out of current at the same 33 Hz, even if I bias 15 mA higher. I don't understand why, as there should be more room until cutoff,....

Not able to understand why, I finally, I removed the 0.075 mm gap material, to achive the minimal gap posible (E and I lamination touching each other). The primary inductance raised by almost 50%.

Now I can see the saturation frequency at 20 Hz (attached). The tube does not get into cutoff region. The low -3dB due to primary inductance is even lower, so this seems the best setup for the OPT. No gap at all (the bias current is just 40 mA).

So the problem did seem to be a too low primary inductance, loading the tube too much.
 

Attachments

  • Fsat_20hz.png
    Fsat_20hz.png
    4.4 KB · Views: 290
The main test that can clarify the performances is the THD vs Frequencies. Normally the shape of the diagram is like a big U, where the THD increase at low and high frequencies. If the shape is narrow the quality is low, if it is wide the quality is better. At low freq. if you increase the dc current the values changes, normally, not so much.

Walter
 
In all the amp I have tested I found that the ones with a wrong performances on THD vs frequency the sound is normally without a energic bass and an harsh treble. This is not related to the power ( in a certain way) but with the current (clean) that the tube can delivery to the load.
One great example is the Audio Note Kit 1.

Here you can find some info about the test done OPT Characterization

Walter
 
Here you can find some info about the test done OPT Characterization
Walter
Interesting thread. I even learned something from it :)

I've taken some measurements of my OPT, aiming to a 1.25 W output, using a SET output. Saturation and roll-off frequency due to primary inductance are below 20 Hz, but I get another issue at a higher frequency, degrading the performance.

At 1kHz, the maximum output before clipping is about 8.8 Vp-p. If, keeping the same input signal level, I go down to 50 Hz, the output tube just enters cutoff, adding some distortion.

Attached, max output at 1 kHz and 40 Hz, clearly showing a portion of the wave, during which the tube is cut (no anode current). Blue is output voltage. Yellow is proportional to anode current.
Magenta is just anode-gnd voltage.

If I add some gap, reducing primary inductance, this cutoff frequency happends at about 20 Hz higher. So, I think it is due to low primary inductance, "rotating" the load line due to the reactance loading. But I measured about 14 H, so it is not really that low, for this low power OPT, is it?
 

Attachments

  • Vmax_1khz.JPG
    Vmax_1khz.JPG
    186.7 KB · Views: 149
  • Same_Vin_40Hz.JPG
    Same_Vin_40Hz.JPG
    180.3 KB · Views: 144
Last edited:
Yes. It was measured in the amplifier. Clearly below clipping. At 40 and 50 Hz.
No, I meant:
How did you measure this 14H value of inductance, at which DC bias, which AC level/frequency and which gap?

Because iron core permeability varies with induction level and that inductance is proportional to the squared turns multiplied by permeability, the inductance also varies with AC level and DC bias. For M6X lams the permeability is low (about 100) at low induction and increases up to about 15000 for 0,8 Tesla then goes back quickly above 1.7 Tesla where the iron is said "saturated".
 
Cathode resistor not bypassed means that the impedance of the tube circuit is higher and this is a problem. And it acts as a local feedback. You must use a fixed bias with a 1 ohm resistor on cathode to check the value of the bias, in this way the coupling with the OT is effectively with the Zout ( Rp that is changing with the value of the bias current ) of the tube.

Walter
 
I did that (using a 5 ohm resistor) with the previous amplifier/OPT. But I wanted this one to be cathode biased, not use fixed bias, and measure how the OPT works in this situation. I left the bypass capacitor out, as I thought it would affect bias if the stage is driven hard. I mean: the bias point is not symmetrical in the load line. Saturation is reached before cutoff. If a high enough signal drives the output tube close to Vgk=0, then the cathode will get lower in voltage, so changing the bias.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.