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 » Sun May 25, 2014 10:38 pm

Image
Image

Slightly over voltage but not so far over as to worry with cheap valves like these.

No grounding caps. Now there is ringing, look at the jagged shape of the lines. This makes a fairly loud mechanical noise, no't know if it's inthe valves or the output torroidal transformer. I hope that will go when the filaments are earthed through the circuit. we'll have to wait. I hope parasitic oscillation doesn't fry up the transformers in the circuit. I am much happier now that this experiment is happening with cheap valves.

Now the test amp is going to be triode 1619 IT triode 1619 IT 6b4g, on a block of wood.
"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 » Mon May 26, 2014 8:02 am

Paul Barker wrote: Slightly over voltage but not so far over as to worry with cheap valves like these.
If a set of 1619 (2.5V) and 6B4G (6.3V) was connected as load, the voltages are very slightly over. A 5% limit would mean 2.625 for the 1619, and I guess 5% is OK with oxide coated filaments - 10% would be OK for the indirectly heated valves.
Paul Barker wrote: No grounding caps. Now there is ringing, look at the jagged shape of the lines. This makes a fairly loud mechanical noise, no't know if it's inthe valves or the output torroidal transformer. I hope that will go when the filaments are earthed through the circuit.
I am surprised at the change of waveform (ringing)?! Nevertheless, the factors have not changed, they are still between 2.28 and 2.22 (i.e. P-P/RMS = factor). Such low factors are indication of "lighter" loading (the 50W load on your unit has a factor of 2.42, while decreasing it to 25W with parallel bulbs decreased that to 2.22).

The mechanical noise happens when the HF AC units (electronic transformers) are not loaded, or not loaded enough. Starting the units without load could damage or destroy them, but later generations do have some "failsafe" mechanism that keeps them safe in no-load conditions, just like the other faulty conditions like "overload" (overheating), short circuit... etc. Thus when you power the unit without load it should exhibit a mechanical noise (vibration) - most probably your units are not loaded enough (and hence the ringing).

Try adding something to draw current, like 5W automotive bulbs or similar stuff. The mechanical noise should disappear, as well as the ringing in the waveform. If the test was done with all 3 tubes connected, that would mean slightly below 20W of power: but I guess you tried with just 2 tubes, and that is just above 12W... too little for the rating of the unit but probably enough to start the oscillation.
Paul Barker wrote: Now the test amp is going to be triode 1619 IT triode 1619 IT 6b4g, on a block of wood.
A 3 stage amp is a good idea when trying to estimate the audibility of HF AC in extreme condition real life amplifiers. On the other hand, it's surreal (for me) to see RH-TTA output tubes used in such manner...

3 stage amps with DH tubes are rare... if it works with this one, there is no reason why it wouldn't work in 2 stage amps, or "classic" amps where only the output is direct heated...

What I currently know for sure is that it works amazingly well with high-voltage-high-current applications like the 813 (and would work equally well with 211 or 845, speaking of 10V heaters). It is important to load the circuit enough - to the point that in extreme conditions low power automotive bulbs or similar should be used for stability - which is something you do not need when a single tube consumes 30-50W of power.

Looking forward to hear the results!

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

Post by Paul Barker » Mon May 26, 2014 10:41 am

That result was with all three valves but that is only 16.3 watts. So that could be the reason for the noise.

So we'll give it a chance. I thought of a lot of ways to raise the demand, but in the end decided to heat an additional 6em7 which puts me up above 20 watts.
"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 » Tue May 27, 2014 5:24 am

Any news? I guess you are preparing the board...

Maybe an automotive bulb would be easier to use than 6EM7? A 10W bulb, perhaps?

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

Post by Paul Barker » Tue May 27, 2014 12:24 pm

Trust me 6em7 easier for me. Anyway have a use for it.

Long pause while a whole Amp is bread boarded.

Patience
"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 » Thu May 29, 2014 12:34 pm

Patience is not a problem... I was considering making one "multi secondary" unit myself, for use with the RH300B - but have not find the adequate electronic transformer near home. I was considering the Philips unit like shown in the video you linked to.

Since some measurements were shown, it would be a starting point, and it does have a toroid output transformer which should be notoriously easy to "rewind" with 3x 6.3V secondaries (2 for use with 300B or 2E22, 1 for use with the driver tubes)...

Meanwhile, I have noticed that a 220uF/400V cap costs slightly less than an electronic transformer, and have calculated that a few 10uF/400V caps plus an LM317 and a few resistors might cost less, while providing more 100Hz ripple rejection, as well as a potential possibility to fine tune the output by tuning the DC voltage at this point.

The two ends of the bridge would have to be disconnected from the board (easy with 1N4007) and connected to the cap/LM317/resistors/cap combo, the voltage than returning to the board at the point where the diodes were leading... obviously this means taking the electronic transformer out of its box... which adds the possiblity to heatsink the transistors...

All in all, I prefer the "added cap + modified secondary" approach... and what I am actually patiently waiting for is your impression on the audibility of the HF AC filament supply, in particular when applied to input/driver tube (direct heated, obviously) positions...

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

Post by Paul Barker » Thu May 29, 2014 1:57 pm

Yes sorry I was distracted by solid state amp project.

Have very full on work Saturday so if I don't find time tomorrow we're looking at second half next week.

But I'm building entire Amp from scratch so not five minutes.
"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 » Thu May 29, 2014 2:33 pm

No problem, take your time. The validity of the results is more important than speed.

The only thing that gets to me is not being able to measure the voltage value accurately. It starts with some value, gradually setting for something acceptable, while the brightness of the filaments doesn't change. I know that it is between 10 and 10.5V, because of brightness, and because of the loss of power experienced with 0.11 ohm resistors (0.55V less). I do not like it with the resistor in series, thus it must be lower than 9.5V with the resistor. While I am getting used to not checking the value, it is slightly irritating to be without means of control.

Interestingly enough, the DC value (measured with the described diode cap adapter), remains unchanged all the time, thus the AC reading is the only elusive detail. I confort myself with a simple calculation: 11.5V unmodified, 1.49x increase with cap, 1.6x less turns ( 5 out of 8 turns ) = 10.7V ... absolute maximum. Unmodified could have been only less at 50+W, 1.49 is the theoretical increase (could only be less), and only 5 turns in the secondary remain of the original 8... it is rather possible that the result is exactly 10V, and falls below 9.5V with the resistor, and the loss of power can be felt.
11.5×1.4÷1.6 = 10.0625, if the increase is for instance "just" 40%... the 0.06V are easily lost in the wire at 5A!

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

Post by Alex Kitic » Mon Jun 02, 2014 1:47 pm

Paul,

regarding the couple of units that you basically rendered unusable after failing to modify the "choke"...

It seems that similar/identical "chokes" can be found in CFL bulbs!
Check out this link: http://www.radiolocman.com/shem/schemat ... l?di=56463

It is interesting that the same oscillator circuit (half-bridge) is used with both ballasts and electronic transformers.

Image

Basically, the "choke" is not a choke at all, rather a transformer with one primary and two secondaries - look at it as a single-ended to balanced converter. The result is two phases necessary for the operation of the two transistors!

This transformer is in direct relation to the minimum current necessary for the operation of the circuit (the primary has to have some current across in order to drive the transistors through the secondaries)... but it has no influence (basically) to the frequency of oscillation. This frequency is actually controlled by an RC network at the input to the transistor gates and the transistors.

Thus you could extract the transformer from a (broken, perhaps, to avoid waisting a good one) CFL. If wired correctly to the circuit, it should work, and it would be interesting to find out how would that change (eventually) the output voltage and/or minimum current necessary to operate the circuit. Every CFL seems to have it, regardless of the oscillation starting mechanism!

Another good news is that the capacitor we are adding at the input is harmless to the operation of the electronic transformer, it maintains a DC voltage level that allows continuous oscillation without help by the DIAC. As I already stated, the peak voltage (relevant to the primary of the output transformer) is already there, thus the transformer is not under any additional stress...

User avatar
rowuk
Old Hand
Posts: 296
Joined: Sun Jun 01, 2014 2:50 pm
Location: Germany

Lower power Supplies?

Post by rowuk » Mon Jun 02, 2014 4:39 pm

Alex Kitic wrote:Paul,

regarding the couple of units that you basically rendered unusable after failing to modify the "choke"...

It seems that similar/identical "chokes" can be found in CFL bulbs!


It is interesting that the same oscillator circuit (half-bridge) is used with both ballasts and electronic transformers.

Basically, the "choke" is not a choke at all, rather a transformer with one primary and two secondaries - look at it as a single-ended to balanced converter. The result is two phases necessary for the operation of the two transistors!

This transformer is in direct relation to the minimum current necessary for the operation of the circuit (the primary has to have some current across in order to drive the transistors through the secondaries)... but it has no influence (basically) to the frequency of oscillation. This frequency is actually controlled by an RC network at the input to the transistor gates and the transistors.

Thus you could extract the transformer from a (broken, perhaps, to avoid waisting a good one) CFL. If wired correctly to the circuit, it should work, and it would be interesting to find out how would that change (eventually) the output voltage and/or minimum current necessary to operate the circuit. Every CFL seems to have it, regardless of the oscillation starting mechanism!

Another good news is that the capacitor we are adding at the input is harmless to the operation of the electronic transformer, it maintains a DC voltage level that allows continuous oscillation without help by the DIAC. As I already stated, the peak voltage (relevant to the primary of the output transformer) is already there, thus the transformer is not under any additional stress...
Does this mean that we can modify these lamp socket converters for filament heating too? That would be cool because they unlike the bigger units don't start at 50-100 Watts......... I can see the E27 or E14 socket on the top of a valve amp now..........

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

Re: Lower power Supplies?

Post by Alex Kitic » Mon Jun 02, 2014 6:34 pm

rowuk wrote: Does this mean that we can modify these lamp socket converters for filament heating too? That would be cool because they unlike the bigger units don't start at 50-100 Watts......... I can see the E27 or E14 socket on the top of a valve amp now..........
Not exactly: while the basic circuitry of most ballasts for fluorescent lamps and electronic transformers for halogen lights share the half wave oscillator circuit, the ballast misses an important part, the output transformer. While this might be rather easy to wind at home provided the adequate ferrite core and lacquered wire are at hand, this goes astray from the concept of easy and cheap sourcing, and light modifications that are relatively standardized. Of course, there would be no need for the socket in the end.

While I am by now quite convinced that we can easily apply the concept to higher voltage higher current tubes like 813, GK71, or 211, GM70 - the multiple secondary concept Paul is experimenting with is more difficult to achieve with predictable results, due to low current draw.

I am also waiting to hear how far can this concept be applied - for instance to direct heated driver tubes.

User avatar
pre65
Amstrad Tower of Power
Posts: 17923
Joined: Wed Aug 22, 2007 11:13 pm
Location: North Essex

Re: Lower power Supplies?

Post by pre65 » Mon Jun 02, 2014 6:41 pm

Alex Kitic wrote:
While I am by now quite convinced that we can easily apply the concept to higher voltage higher current tubes like 813, GK71, or 211, GM70

I am also waiting to hear how far can this concept be applied - for instance to direct heated driver tubes.
Those four valves you mention ARE directly heated. :?
Life is what happens to you while you're busy making other plans. John Lennon

G-Popz THE easy listening connoisseur. (Philip)

User avatar
IslandPink
Oooo, Roberts Radio's sound nice
Posts: 7427
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

Post by IslandPink » Mon Jun 02, 2014 6:57 pm

Driver tubes, Phil :oops:
"The bass is the king of the instruments - it has no known natural predator" (Wobble)

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

Post by Paul Barker » Mon Jun 02, 2014 7:02 pm

Might get away with a hybrid of the two, lift transformer from 20 watt halogen supply and bolt it on the fluerescent lamp supply.

Why would you want to do that?

To get below the 20 watt level, but I suppose a dummy load on the 20 watter is a better adaptation.

I don;t think I'll fiddle with the broken ones thanks Alex. Not enough to gain for the effort considering the time constraints in my life these next few weeks..
"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 » Mon Jun 02, 2014 7:35 pm

What I meant with "driver tubes" reflects 2 issues:

1) normally you would use a lower power tube as driver, like a 10Y or 45 or 6B4G, where we hit the "minimum current draw, or power", since a 6B4G consumes just 6.3W, and even for both channels that is way lower than 20W;

2) while you might adopt the Sakuma style and use a GM70 to drive a GM70 (60W power in the heater, or even more, actually), what you do with a driver or input tube gets further amplified down the chain: including hum/noise/buzz.

Actually, the 2nd point is what I am most curious about. The 1st can be solved, if not more elegantly, by adding a parasitic consumer (like an automotive bulb or similar).

Post Reply