HF AC DHT heating

For the three and more legged things
User avatar
pre65
Amstrad Tower of Power
Posts: 21373
Joined: Wed Aug 22, 2007 11:13 pm
Location: North Essex/Suffolk border.

#76

Post by pre65 »

I just looked at the ones on my 833a amp, and there are no markings on them. :(
The only thing necessary for the triumph of evil is for good men to do nothing.

Edmund Burke

G-Popz THE easy listening connoisseur. (Philip)
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#77

Post by Paul Barker »

Alex Kitic wrote:
BTW, what size is the ferrite core that you are using in your multiple-secondary project?
Image
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
pre65
Amstrad Tower of Power
Posts: 21373
Joined: Wed Aug 22, 2007 11:13 pm
Location: North Essex/Suffolk border.

#78

Post by pre65 »

Alex Kitic wrote: I was also considering the use of NTC devices in series to limit the current inrush. Any experience with that?
Alex, after a bit of searching I reckon these are what I'm using.

http://www.rapidonline.com/Electronic-C ... 5a-26-7608
The only thing necessary for the triumph of evil is for good men to do nothing.

Edmund Burke

G-Popz THE easy listening connoisseur. (Philip)
Alex Kitic
Old Hand
Posts: 369
Joined: Mon Oct 14, 2013 3:20 pm

#79

Post by Alex Kitic »

pre65 wrote:
Alex Kitic wrote: I was also considering the use of NTC devices in series to limit the current inrush. Any experience with that?
Alex, after a bit of searching I reckon these are what I'm using.

http://www.rapidonline.com/Electronic-C ... 5a-26-7608
Assuming that you are placing it in series with the primary, what is the current draw in operation (once stabilized)?

I was thinking of 16 ohms 2.9A for the electronic transformer only, and maybe 10 ohms 3A for the main toroid (300VA consumption, 400VA specified, on 500VA core).
Alex Kitic
Old Hand
Posts: 369
Joined: Mon Oct 14, 2013 3:20 pm

#80

Post by Alex Kitic »

I did some listening to the RH813 yesterday evening, and while i was not particularly in the mood for serious evaluation, I did not notice any relevant difference between using separate units and the 2-secondary windings unit.

1) Voltage is obviously adequate, probably around 9.7V - and compared with the probably around 10.3V of the separate units, there was no noticeable difference in sound. Strange enough, since I recall not being particularly happy with the separate units and added 0.11 resistor (0.55V less). Maybe this is some placebo effect?

Anyway, if there is no difference in sound, even 9.5V is better than 10.5V, to mention extremes.

2) Hum/noise/buzz - inaudible and drained in the tube hiss of the preamp. Without preamp, listening with one ear literally on the woofer or midrange, there is maybe very slight buzz, meaning slightly more than with the separate units (who can say whether this is correct? This is something one needs a scope for - detecting the otherwise inaudible). This is obviously due to the increase in current vs same value cap. While it might be improved with a regulator filtering off the ripple, it is totally irrelevant in real world operation since whatever there might be gets totally covered when the preamp is powed on.

How does this relate to having a high efficiency speaker system? Well, I guess if the tube hiss was to prove audible at all, one would first have to change tubes... to be able to hear anything else. Than again, people usually listen to music from their listening spot, instead of putting an ear in the woofer to listen to noises....

The most important fact is that there are no audible high frequency artifacts, which might creep in because there are no filters at the output. This unit is specified as 30kHz, and while I still believe we cannot hear that, due to the hummer resistors we should be having only residual 2nd harmonic (60kHz) and higher, which I expect cannot be detected even by scope at the output of the amplifier (secondary of the OPTs, bandwidth issues).

3) Operational issues - none. No blown fuses, although I intend to add an NTC at the mains input of the electronic transformer, and probably another at the input of the rather powerful toroid that supplies HT, rectifier, and driver tubes heaters.

The unit runs cold, and it is my impression that it runs colder than the other units (measuring the temperature of the original unit housing. Obviously the small heatsinks on the transistors help keeping it cool.

I am satisfied with the result, and actually not particularly motivated to try regulating the rectified DC that supplies the half wave circuit. If I decide to do it, it will be in the interest of science :)
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#81

Post by Paul Barker »

Great news Alex.

My multiple tap additional transformer is a resounding success.

The voltages as averaged by the scope are 6v and 2v, and that is exactly what I intended. The fundamental frequency is 53khz. The power draw is 45 watts from a 20 to 60 watt transformer.

Image

To try and help you get a handle on what a normal voltmeter would read, I dug out an old Fluke which doesn't say True RMS on it. But it seems that it probably is true rms.

Image

Once again we have a discrepancy between the scope averaging and a multimeter. If you recall my bench multimeter was also reading a greater voltage than the scope.

At least in my case if the multimeters are right I am safe, if the scope is right I am not enough under voltage to worry about.

It seems like most of this thread we are talking about voltage.

Only the 6em7's give a clue they are lit, but they look just right, if that helps.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#82

Post by Paul Barker »

Acording to reply number 2 on this thread the only version of this model which is true rms is the 27/FM.

but the guy may not know, or there may have been an upgrade to this particular one, or the figure it reads can't be relied on.

so we are back trusting our valves mostly to the scope.

these specificationsindicate 4% accuracy up to 30khz. Obviously it works at 52khz but we have no idea about acuracy.

I am back to trusting more in the scope.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Alex Kitic
Old Hand
Posts: 369
Joined: Mon Oct 14, 2013 3:20 pm

#83

Post by Alex Kitic »

Great news indeed! I assume you could get multiples of a value, like 1.1 volt per turn - 2.2V, and 6.6V. A further check of your measurements could be multiplying the 2.5V equivalent value 3x.

For the RH300B, I plan to try getting 3 identical secondary windings at approximately 6.3V - that is much easier. Assuming 6.3V, to get the 5V I would use a 0.56 ohm resistor per heater leg.

Now I am looking forward to hearing your impressions of noise/hum/buzz and eventual artifacts. The real test is with the driver tubes. My expectations are high, but if you'll get to hear anything spurious it will be 200Hz, i.e. 2nd harmonic of 100Hz. Thus the key to total inaudibility is ripple reduction after the rectification. The trick with regulators is the way to do it cheap and efficient, but the regulator will need a nice heatsink: in your case at 20W this means 0.087W per V of drop across the regulator, but in my case it's 0.44W per V of drop, and if you assume 5V, that is above 2W. In your case you might get away without hearsink or a very small one, depending on the ripple after rectification.

******

I really envy your resources - when you find an old DMM, it's a great old Fluke. Of course, the scope is more accurate, because all DMMs increase in error with increase of frequency. My DMM measures differently 3 units that are quite different, probably in frequency. The very cheap unit I mentioned, simple and without frills, is measured in original form at 12.8V. That is probably a very low error, if we assume the DMM is measuring the peak value, which is probably to be factored by 1.1 in respect to a square wavw where average and RMS are identical - most probably because the frequency is quite low. The other 2 units are about 26V in original form, the error being probably very high. No accurate prediction possible. Your old Fluke looks perfectly precise in comparison.

I was thinking about using some pc scope software, but the question is whether the sound cards can register such high frequency at all, and the need to use resistors to divide accurately (?) to acceptable levels (the sound card cannot take 10V signals, thus a voltage divider for a 10th or 20th is necessary.

Without a scope, we are left only with the mathematical calculation.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#84

Post by Paul Barker »

If you have a x10 probe you have your 1/10th division.

Yes the old 25 is a thing of beauty. The case is also water tight and double skinned so that it floats.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Alex Kitic
Old Hand
Posts: 369
Joined: Mon Oct 14, 2013 3:20 pm

#85

Post by Alex Kitic »

Paul Barker wrote: If you have a x10 probe you have your 1/10th division.
I was referring to a sound-card based scope for the PC. This implies "no probes", rather connecting the signal to the sound-card line input.

Since the input cannot take more than 1V, and I intend to measure 10V, I should prepare a voltage divider (like a wirewound pot) to use as a means of reducing the signal. For instance, I leave just 2.5% of the signal, getting 25mV per 1V of input (safety and feasibility).

The other problem I anticipate is frequency - I need to assess a signal which is higher in frequency than 20kHz (30-50k in most cases). A normal soundcard is able to go probably up to 24kHz (48k sampling ratio).

Well, I did download a very nice little software that seems to have enough features for my needs, and seems to be simple to use. But, just as I anticipated, with the sound-card in my notebook it goes only up to 24k if I choose the 48k sampling option.

I might try it in one of my classic PCs, wtih sound-cards that can record at 96k, or a maximum of 48kHz signal input...

Another issue I always mention when measurements made with soundcards - is the noise of the PC/card ... attached a nice capture from the software: yes, it seems that other than frequency limitations it is able to do what it takes to measure and estimate what I need?

Any advice on the voltage divider? Would a wirewound 100ohm pot do, if connected to the output of the transformer like I would connect a DMM (of course, 100 ohms to the output of the transformer, scraper and one and, at some pre-set and measured value giving a known and adequate voltage division)? Would I than have just to multiply the measurement with the voltage division factor?
Attachments
scope soudcard only_bw.jpg
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#86

Post by Paul Barker »

No, I would still use the X10 probe because it is ideal in the situation, up to 100khz it would need no extra networks to be acurate.

the x10 probe has a 9meg series resistor and a cap parallel with thatt 10 to 12pf. then there are some compensation networks depending on how much you spent on your probe. If only one it is just a variable zobel network with a shunt resistor and trimmable cap to ground. This is adjusted to make the callibration squasre wave flat at the leeding edge.

Then at the scope there is a shunt resistance of 1meg. This makes the voltage division. On the X10 probe you can select it off so you then go straight into scope where the 1meg shunt waiting and you get the entire amplitude.

You have to find out the input resistance of you sound card, make a series resistor to devide the signal down with this. Or if possible alter it to 1meg and buy a cheap X10 probe and adapt it to your sound card.

I have no idea about soundcards, but you must find out it's input impedance.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#87

Post by Paul Barker »

Let's look at this another way. Measure the input impedance of the sound card and then put your signal in through a series resistor which devides the signal down to a tenth. Won't make a pretty square wave but will give you the RMS aslong as the software has a feature to do so.

One general principle is that a pure square wave the RMS is near enough the peak voltage. In our case we have a sloaping horizontal line but I recon measure the peak from the centre of that sloap and you have your RMS.

but for an absolute certainty the way to be completely sure of your voltage is to use your 317 to control voltage. Now get a true AC at 50hz that is exact and heat your valve. MEasure your anode voltage. Now use your HF supply on same valve all else equal and adjust your 317 controlled voltage so anode voltage is exactly the same. Now you have a strong benchmark. find out at that position what your software says the RMS is and adjust. The power to perfom the work is the true RMS and the totally true way to check that is to make sure it does the work exactly right. No guessing.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
User avatar
Nick
Site Admin
Posts: 15706
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#88

Post by Nick »

Paul Barker wrote:Let's look at this another way. Measure the input impedance of the sound card and then put your signal in through a series resistor which devides the signal down to a tenth. Won't make a pretty square wave but will give you the RMS aslong as the software has a feature to do so.

One general principle is that a pure square wave the RMS is near enough the peak voltage. In our case we have a sloaping horizontal line but I recon measure the peak from the centre of that sloap and you have your RMS.

but for an absolute certainty the way to be completely sure of your voltage is to use your 317 to control voltage. Now get a true AC at 50hz that is exact and heat your valve. MEasure your anode voltage. Now use your HF supply on same valve all else equal and adjust your 317 controlled voltage so anode voltage is exactly the same. Now you have a strong benchmark. find out at that position what your software says the RMS is and adjust. The power to perfom the work is the true RMS and the totally true way to check that is to make sure it does the work exactly right. No guessing.
Before you did that if you wanted to be certain, you would need to see how much the anode voltage varied with the 317 reg voltage varying a known amount (say +- 0.5v), you don't at the moment know if the anode voltage is a valid way of measuring heater voltage.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Paul Barker
Social Sevices have been notified
Posts: 8863
Joined: Mon May 21, 2007 9:42 pm

#89

Post by Paul Barker »

Yes I see what you mean. Know what differences a known change in heater voltage inspires.
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
Alex Kitic
Old Hand
Posts: 369
Joined: Mon Oct 14, 2013 3:20 pm

#90

Post by Alex Kitic »

I am not getting what anode voltage, bias voltage, or current draw has to do with heater voltage, i.e. voltage across the filament?

The LM317 sets a current draw which will be same with any tube under amy operating condition, if the tube can be operated under that condition. The only change with a different tube or even tube type is cathode voltage.

Literally, you could plug in a GM70 and it would draw the same current, but bias voltage as cathode potential would be different.

On the other hand, to get a true RMS reading of the heater AC voltage you do not even need to pass any current through the tube, just the filament matters.

Imagine using your tube as a light bulb (like that janitor). The voltage across the heater may change, but not because all of a sudden you are drawing current through this tube that has been transformed from light bulb into amplifier device.

The procedure you are describing might be applied to a DC on the heaters condition where there is a difference in potential between the ends of the filament and thus may be related to current draw and cathode potential.

I can't see a reason why AC voltage across the heater would change with a change of anode voltage or current draw, or without B+ (take the rectifier out of the socket)?

As for the probe, I haven't got one. I could use a sound card in a PC to view the waveform on the scope application, and eventually measure it. Since this is no scope, but a sound card and software, I need to adapt the input voltage to the input capability of the sound card. And, with a degree of precision so that I can easily calculate the result for the full scale signal.
Post Reply