HF AC DHT heating

For the three and more legged things
User avatar
Paul Barker
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sat May 24, 2014 8:15 pm

Alex Kitic wrote: each secondary needs its own cap. On mine the value cannot be read: what should be used (it's for RF filtering if I'm correct?!
211 pF

Image
"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
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sat May 24, 2014 9:58 pm

I have a problem.

Firstly the lit valves are all oxide and the brightness is about right. Front right is #10 back right is #27 front left is PX4.

From filament brightness they look alright.

BUT the scope is totally confused. There is some interaction between the secondaries and this scope which is automatic hasn't a clue what to latch on to. We have induced interactions between the three taps.

No doubt in terms of audio this would be a total car crash.

So project for using one of these to heat more than one valve has crashed and burned. :cry:

Image

Nothing ventured nothing gained.
"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

Post by Alex Kitic » Sun May 25, 2014 7:34 am

Paul Barker wrote:I have a problem.

Firstly the lit valves are all oxide and the brightness is about right. Front right is #10 back right is #27 front left is PX4.

From filament brightness they look alright.

BUT the scope is totally confused. There is some interaction between the secondaries and this scope which is automatic hasn't a clue what to latch on to. We have induced interactions between the three taps.

No doubt in terms of audio this would be a total car crash.
You are mentioning "taps", but I guess those should be completely separate secondary windings. If it works on a normal transformer for 50Hz, there is no reason why it shouldn't for a HF unit?! There must be some simple explanation.

If we wire 3 light bulbs in series from 3 taps on one secondary, they will all light up nicely with the correct voltage and thus brightness. But if those are filaments, they must be completely separated to avoid feedback.

If there is no continuity between these windings on the transformer (and i assume you are aware of that and have checked it properly), it could happen elsewhere. Remember my first attempts at HF heating where it would not light the tube in circuit, but worked on the tube pins? I had caps bypassing the cathode from each end of the filament that acted as dead shorts! I understood that a classic approach was necessary, bypassing only from the mid point of the hummer pot to avoid shorting the secondary to ground (and being planned for HF AC heating, the RH813 was wired in that manner from the start.

Your short is between secondary windings only, and tubes are lit. But there are caps from one end of each filament to the negative pole of the rectifier/oscillator circuitry (I assume you have installed those?). In AC these caps are shorts between the secondary windings).

Now there are 2 alternative outcomes:

1) The scope is confused, but that is a HF issue and would not be heard at the output of the amplifier. This could be true, but I am not so sure. Depends on the frequency and magnitude.

2) You remove all (or all but one) RF cap. This way there will be no HF shorts (thus feedback) between the secondary windings, i.e. cathodes. RF might or might not become an issue, and would probably be inaudible for humans.

Now, the 27 is indirectly heated, am I right? There is no direct interaction between it's heater and cathode. This leaves the 10 and PX4. I guess you are using the 10 to drive the PX4. If you have an interstage transformer there, they could be in phase (positive feedback! ) ... and the PX4 might be fixed bias (filament at ground potential, in direct connection to ground). I would leave the cap just on the 10, but if the 10 is fixed bias as well the windings would be probably shorted again (PX4 filament is ground and 10 filament is ground potential thus they are directly connected and this applies to being connected with the negative pole of the oscillator in high frequency AC... in that case they are shorted anyway and would light the tubes but could pass the garbage: the system is totally safe for cathode bias, but fixed bias on all tubes might be a problem with more than one secondary... or even separate units).

This was blind guessing about what happened at the crime scene, but I am pretty confident that the experiment is not dead and buried!

Now that I mentioned it, I was powering one 2E22 filament without problems although that is just 1.5A, and at nominal 11.5V it was only about 17W! Of course, in my view it is secondary current that matters most in this equation (modifying the secondary to have 6.3V at the output instead of 11.5V still represents a current draw of 1.5A, thus it would work the same, but without burning excess voltage. It should work for 2x 1.2A heaters (300B) and some additional current for the driver/CF...

User avatar
Paul Barker
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 8:17 am

OK useful post.

Firstly I placed three caps 211pF at the end of each individual secondary and brought all three to a star ground together before taking that to the same place on the board at the original cap.

Right, how it is behaving.

It is frightening me to death because of the value of the valves I am using and the test results are just way out from expectations.

So originally set up the original transformer for 25 watt load, so I could establish turns per volt at that load.

Using that finding 1.27 volts per turn, I wound bifilar the 2.5v because of the extra current draw of this valve, the rest I used one half of the remaining bifilar wind each, cut to length appropriate to heat each other valve.

How does the scope and incidentally the Hameg respond.

1/ the voltage readings are bizarly and unequally (not same proportional problem on each valve) displayed.

The 27 and the 10 are displayed rediculously under voltage I didn't leave it switched on long enough to write down due to not wishing to kill expensive valves. But the 27 was about 1.7v on scope the 10 was 5v on scope the PX4 (this one scared me) was 7v on scope.

Quickly as soon as I saw PX4 result switched off. Raced to check that I had not accidentally transposed the PX4 with the 10. But no I had not.

So switched on again and checked the voltages with the Hameg. IT told a slightly better story. The 27 was the same the 10 was above 6v don't remeber exactly and the PX4 was more like 5 or 5.5v. So things were less of a worry. I took the picture shown having breathed again and left it running long enough for the photo then turned off again.

Just an asside if you don't let smoothing caps discharge it won't turn on again. It needs to start slowly if smoothing caps are up to voltage when you turn on it won't power up. so they need a fast discharge resistor in situ.

Simultaneously on the scope I read rediculous frequency. No I didn't think at the time to attach the frequency counter. It was all done in a panic. It was something like 27khz.

Then I recalled that the 10 and the px4 secondaries run together as they are each half of the bifilar setup. Could the px4 be inducing part of the voltage of the 10 yet without power. in household electrics this occurs if you run low voltage cable with high voltage cable, your meter will detect a voltage when there isn't any power. So with that in mind, unless I attach an ameter, I have only the filament brightness to go on.

The PX4 looks a little bright what do you think?

What is this bizaar frequency?

And what interactions might take place at the caps?

Out of my depth.

OK will pick it up again but not share bifilar runs and will change the valves I experiment on. Maybe go find 12E1's at the the lockup they are next to useless. Two or three of those might be worth testing a new multi seocndary rig on, so I can spend longer with them connected and try to establish the picture in less of a panic.
"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
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 8:25 am

Then after stabilising it on 12e1's I might change the amp for the moment to something like 6j5 to 6b4g to 6b4g

Regarding the topology yes original idea was 27 RC coupled to 10 bifilar wound IT coupled to PX4.

So the test set up would be same.
"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
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 8:38 am

looking at the pictures again, if the PX4 was provided a genuin 6 or 7v with power and the oxide filament 10 a genuine 6v with power and the 27 a genuine 1.7 volt we should see the 27 and the 10 consireably less bright the the px4 considerably more bright.

Though there is a bright spot on the px4, this can occur naturally. I hope that the voltage reading that don't stack up to the turns count are induced voltages which bare npo relation to the voltage that is powering the filament. A thought reinforced by the strange frequency observed. The scope has latched on to what is happening at this frequency in error. If it were an analogue scope I could change that tbut this one just latches on to just one frequency I can't figure out how to usher it away from that and look at another. There probably is a button press menu combination. I just haven't found it.

I have one of those Tectronics scopes the size of a large microwave oven with about 100 ECC88's. I had kept it just to sell the valves in it. IT seems so impractical as a scope.

But looks like I better dust it off, at least I can turn knobs and it stays where I put it instead of boinging back to it's pet frequency in an instant.

I'll have to bring up it's power supply ever so slowly on a variac to reform the caps as it hasn't been used for perhaps 10 years, and anyway it must be 50 years old. It did work fine last time I used it. Just so big and heavy. Looks like I have a project for it now. I bet it won't be phased by whatever is happening here.
"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
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 9:10 am

See the red power light at the bottom.

Giving her 100v to start with. How long do you think she needs at 100v before I raise it a notch?

Image
"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
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 9:53 am

Ok checked the manual it has 19 6DJ8's a bunch of 6AU6's (about 10) a single 12AX7 and a few ECC82's and some other bits and pieces.

Definately worth more in parts. But great to initmidate the police when they think you are making bombs :lol:

It's a strange story. We had three Pakistani lodgers who were builders building the new Sainsbury's. I was out at work, wife gets home to find house surrounded by police and immigration officers. At first going quite well, Pakistanis all three legit, passports and permits checked out. then they saw my man cave. Next thing you know senior officer on radio , nee naw nee naw armed response turns up.

My sweet angel of a wife facing off to half of Yorkshires police force.

Me still driving back from Wakefield (where the real terrorists probably are). anyway

"what's all this electronic stuff (while thinking "bomb making equipment")"

"oh that's what he does? come inside and have a look"

They are shown a bunch of breadboards loads of valves and transformers.

they stood down. never seen them since. They didn't even wait for me to get home.

the seargeant who called in armded response is probably still being ribbed about it back at the station :lol:
"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
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 9:57 am

She may be an old scope, but from memory of using her, she always did as she was told and displayed a readable stable trace. This modern electronic one behaves like a 2 year old who always gets what it wants and won't tell you what you want to know. Until I discover the manual override that is, if there is one.
"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

Post by Alex Kitic » Sun May 25, 2014 10:14 am

1) You need to try it on an amp where at least one tube is direct heated, to be able to assess the effect on sound and hum: 6B4G is an option, 12E1 is indirectly heated (and otherwise a peach since it can be used in my Universal as direct replacement for 6550, just the way I use EL38 as direct replacement for expensive EL34).

2) You seem to be preoccupied with voltage, but by now I am getting used to it and it works rather predictably. Once heated up, voltages should stay stable and the way windings interact is identical to a normal transformer.

3) You did not mention disconnecting the 211pF caps from the windings? With those caps connected, there is heavy interaction between your windings, because they are shorted together (in high frequency domain, and that is where it matters with this transformer).

4) Last point: is any of your direct heated tubes operated as fixed bias? If it is, one end of the filament (at least is grounded). If both direct heated tubes are fixed bias, their filaments are shorted by connection to ground!

The valve scope is impressive and I envy you for having such equipment at hand... all I have is a garden variety DMM, simulation software, and a calculator. But the scope will not help if the secondary windings are shorted together either in high frequency domain, or totally by connection to ground for fixed bias operation.

The first sign of this will be voltage instability. So disconnect the caps - but if fixed bias, you will need either to test on a different circuit, or modify at least one to cathode bias.

Nevertheless, the valve scope will be a gret way to check on things ;)

User avatar
Nick
Site Admin
Posts: 12127
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Sun May 25, 2014 11:05 am

Just a question. Do you know if the 211pf cap is to ground, or is it feedback for the driving circuit?
Resistance isn't futile it's V / I.

User avatar
Paul Barker
Oooo, Roberts Radio's sound nice
Posts: 7032
Joined: Mon May 21, 2007 9:42 pm

Post by Paul Barker » Sun May 25, 2014 11:56 am

It may be feedback. But the radio ham in the YouTube I put up thinks it grounds througb the power supply.

The units are class I I so not earthed.

But definitely something happens due to the caps.

I'm changing valves next before continuing.

But off to in-laws soon.
"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

Post by Alex Kitic » Sun May 25, 2014 11:58 am

Nick wrote:Just a question. Do you know if the 211pf cap is to ground, or is it feedback for the driving circuit?
The 211 pF cap - actually, a ceramic type cap Paul has measured as 211pF - is not a feedback device and can be disconnected.

On my unit you can see a similar cap marked as C7 (picture attached) and it is connected between one end of the output and the "ground" of the oscillator circuit, i.e. the negative side of the bridge.

The only purpose of this cap is RF filtering. I have found a generic schematics that is illustration for our purposes (and strangely enough the cap is marked C7 as well).
Image

If we look at the secondary winding and cap as an RC filter, I estimate it would be -6dB somewhere around 60kHz or more... but if we connect several secondary windings to the same point via these caps, they are shorted together at high frequency - which makes them interact.

It is safe to disconnect the cap, although it might increase high frequency output to the filament. My point is that this high frequency content will be efficiently enough filtered by the amplifier circuit and output transformer, and we will not be able to hear it anyway (it should not make much of a difference in practice even if we were to remove it from units with only one secondary winding output).
Attachments
IMG_2404_web.jpg

Alex Kitic
Old Hand
Posts: 369
Joined: Mon Oct 14, 2013 3:20 pm

Post by Alex Kitic » Sun May 25, 2014 12:08 pm

Paul Barker wrote:It may be feedback. But the radio ham in the YouTube I put up thinks it grounds througb the power supply.
Indeed, it grounds the output to the power supply of the unit (i.e. connects to the same point where the negative of the bridge, and the cap we are adding across the bridge, connects).

But it is not any means of feedback - just RF filtering!
Paul Barker wrote: But off to in-laws soon.
And I hoped you were somehow freed from that obligation ;)
("off the hook", I guess that is the expression)
Like, your wife saw you were in dear trouble, so she decided to leave you home alone to sort your problem out :D

As soon as you remove the caps, you will sort out most of the problem, and you will be able to get "normal" frequency readings as well.
Have you tried the secondaries each with a separate light bulb (automotive 5W "position" bulbs should work fine for that and are dirt cheap)? If it works for them, the only reason it would not work fine for tubes is that there is something "shorting" the secondaries together...

User avatar
Nick
Site Admin
Posts: 12127
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Sun May 25, 2014 12:12 pm

Ok, fair enough. But one last thing.
My point is that this high frequency content will be efficiently enough filtered by the amplifier circuit and output transformer, and we will not be able to hear it anyway
Consider the Dalek
Resistance isn't futile it's V / I.

Post Reply