Heater Supply Circuits-AC and DC

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
When you ran fluorescent light tubes from DC (not a common thing), you got better life if you swapped the ends occasionally.

Not applicable to vacuum tube heaters.
 
Switched Mode Power Supplies (SMPS) are much better than they used to be, with switching frequencies well above the audio band. The addition of hiccup mode short circuit protection designed to fire up into large capacitive loads means they are now suitable for consideration as heater supplies. I have started using them myself in tube lunch box projects that require 12V at up to 3 amps. Inrush current is still an issue so I overate the SMPS so for 12V @ 3A I will use s 100W rated unit

My favourites right now are mad by MeanWell. Despite their origins they are are sizeable and well respected company and often used in telephony/GSM operations where longevity and reliability are paramount. At present, the one I use most is the MeanWell LRS-100-12.  It is available at low cost from TME:

https://www.tme.eu/gb/details/lrs-100-12/industrial-power-supplies/mean-well/?gclid=EAIaIQobChMIk9qMhun61wIVSrvtCh0w0AOXEAAYAiAAEgIPLfD_BwE

Datasheet is here:

https://www.tme.eu/gb/Document/5adfc6ee9376b871796d0b1e9b29d2ad/LRS-100-SPEC.pdf


I will add other types as I test them.

Cheers

Ian
 
saint gillis said:
A friend of mine told me that DC in the heaters could after a long period of time polarize the heaters which was not good, and he said it was good to have a switch to reverse the 2 heaters supply pins time to time , to avoid heaters polarization, what do you  think of that?
Hum can noticeably increase in some input stages if the heater-to-cathode leakage resistance falls (it's typically many many megohm) - which is usually fixed by tube-rolling.  The heater to cathode interface could degrade differently over time if a DC heater was used - perhaps due to metal dendrites or other changes in the alumina baked insulation.  So sounds plausible to me.
 
trobbins said:
Hum can noticeably increase in some input stages if the heater-to-cathode leakage resistance falls (it's typically many many megohm) - which is usually fixed by tube-rolling.  The heater to cathode interface could degrade differently over time if a DC heater was used - perhaps due to metal dendrites or other changes in the alumina baked insulation.  So sounds plausible to me.

I have often come across poor hk leakage especially in NOS tubes. I am pretty certain the vast majority of them were operated on ac heaters so I see no evidence dc heaters are to blame.  Most of my designs require elevated heaters and they very soon show up poor hk leakage in a tube, I am sceptical of dc being a cause. Pro audio mixers from the 40s often used dc heaters. Those guys knew what they were doing, I have all tube mixers with customers that have been operating for several years on dc heaters and I have never had a complaint about hum.

Cheers

Ian
 
There are many  degradation mechanisms of hk resistance it seems, and many are due to quality and variation in the heater insulation coating across manufacturers and years of production. So I wouldn't exclude it.
 
trobbins said:
There are many  degradation mechanisms of hk resistance it seems, and many are due to quality and variation in the heater insulation coating across manufacturers and years of production. So I wouldn't exclude it.

I won't exclude it but neither will it be anywhere near the top of my list.

Cheers

Ian
 
ruffrecords said:
Pro audio mixers from the 40s often used dc heaters.

This I have never seen.  Love to have an example pointed out.  Only thing I can think of is 1st stage gain reduction tubes.  Literature of the era points out that, given capacitor size/quality of the time, DC filaments can often have more hum than a proper AC filament. 
 
emrr said:
This I have never seen.  Love to have an example pointed out.  Only thing I can think of is 1st stage gain reduction tubes.  Literature of the era points out that, given capacitor size/quality of the time, DC filaments can often have more hum than a proper AC filament.

Altec 250 SU. Its 535A power supply provides 6.3V DC at 13 amps. Peak to peak ripple at 13A is 1.5V.

Cheers

Ian
 
ruffrecords said:
Altec 250 SU. Its 535A power supply provides 6.3V DC at 13 amps. Peak to peak ripple at 13A is 1.5V.

Cheers

Ian

1960's, and probably the only commercially produced American console I can think of that did. 

Now, up until about 1940 they may have used batteries for filament.
 
ruffrecords said:
Most of my designs require elevated heaters and they very soon show up poor hk leakage in a tube, I am sceptical of dc being a cause.

Ian, you would be elevating the heater to avoid stressing the hk voltage beyond rated limits.  The hk insulation in that situation is then stressed towards but hopefully under rated limits due to signal swing, and the likely large elevation levels.  Can you elaborate on how you then distinguish a 'poor' valve - is it from a constant high level of mains frequency leakage current (due to the heater AC waveform), or an ad-hoc pop-corny style breakdown of that insulation, or some other observation?

It's also worth noting that many use a Vdc elevated heater voltage (to a much lower elevation level, where hk voltage is not stressing the insulation) to minimise hk leakage resistance - which results from an interesting variation of hk resistance when hk voltage passes through 0V.

As a general comment, its much easier nowadays to distinguish mains rectifier related noise artifacts in audio signals due to modern spectrum analyser software and very low noise floor adc equipment/soundcards.  To observe those hum artefacts back in the 50-60's, the hum level would need to be significant, and I guess the S/N of recording media would inherently mask hum from earlier in the signal processing path.
 
Speaking of unusual heater circuits ...

I've just overhauled a late 60s japanese stereo integrated tube amp with pp el84 finals..

The heater supply for the phono preamp section is the cathode current return from the finals .. and thence to providing bias at the finals' grids !

ie.    el84 cathodes -> some resistance  -> phono section AX7 heaters (series heaters) -> el84 grids

It all works out well - the EL84 cathodes are around +20V @ 60mA for each pp pair - the stereo pp pairs are connected together, after some rheostat adjustment, giving a total of 120mA current at +20V or so.

That feeds a pair  AX7s with dc (under) heating (series tubes with series heaters) for stereo phono preamping....

with around +10V@120mA for each AX7 tube then to ground  ... *also* then with a connection of the second AX7 heater  (+10Vdc) parallel sent to the grids of the pp EL84 stereo pair  ... as a 'fixed' bias  (each with 250K resistances grid to ground)

Quite roundabout, for sure -  but one does wind up with combo fixed/self bias on the EL84s of around +11Vdc cathode-to-grid  *and*  dc heaters for each of the phono AX7s  of +10Vdc at 120mA.

I was amazed but it does make sense when one one wants underpowered dc heaters on phono preamp section without the cost of an additional heater winding on the traffo  plus rectifying/filtering  :)

The el84s are thus biased very hot, at +11Vdc cathode-to-grid, tending towards the class A pp mode - not too efficient, resulting in some 10Wrms into 8ohms each channel at 3% THD.

It surely does sound good - at lower power of 5Wrms or so, THD drops to around 0.5%.


Those crazy 60s Japanese hifi tube designer gurus for low cost solutions!

...

I recapped where necessary, rebiased for lower sensitivity on the aux inputs and got the hell out!

It was way too sensitive to use with a modern computer audio interface, so I set  up the 'low sensitivity'  aux inputs  to handle around 8Vpp for max power ...  I left the tuner/tape and phono mag/crystal, mic, tapehd  inputs as they were.

One could rework that whole 'high gain' section to be a killer guitar pre, but I resisted  and left the old RIAA, NIB weird feedback eq'ing as is  :)

This  'Encel' el84 amp featured an elaborate speaker-with-feedback-coil capability, called  'Motion Feed Back' and a complicated 'Damping' control to boot .....  I replaced all that with a simple variable nfb circuit.

Replaced the original tube set  for some fresh ehAY7 (preamp),  jj5751 (cathodyne) and russkie 6P14P ev finals ....

The amp was running the original australian made 'milliwatt el84' finals at some high supply  ..  like +370V plate, +320V screens  :eek:   

Amazingly, these tubes, circa late 60s tested very well!

I replaced with some nos 70s 6P14P ev  ..  which are a good sounding and cost effective rugged el84 sub ...  they are rated for 400V plate, 300V screen with 14W max plate dissipation .. a bit of a 'super el84/6bq5'.

I dropped the screen supply to 300V at first, but then returned it to 320V ...  for a little more power at 3%  ....  the robust russkie has a generous screen dissipation and seems to have no trouble with the higher than spec supply. 

The bias setup doesn't lend itself to tinkering  :mad:  I tried with no success  .. always seemed to 'equilibriate' back to +11Vdc cathode-grid but with higher THD, so I gave up!

...

Happy with the result - it's equally at home with low and high senstivity inputs with a lot of adjustability in between.

I'm using it as a stereo FX amp with hifi speakers, but it really sounded good and loud! on a 12" jensen alnico cab too,  with a DI box on some guitar  8)

It's super quiet ... absolutely no hum even at full power  ..  about the best hum performance I've come across in a setup with significant gain - it has those old  style steel alloy close fit shield/heatsinks on the phono AX7s.

Quite a feat really, considering it has a voltage doubling HV supply circuit, which usually give more ripple, *and* the ancient main resevoir caps showed no sign of degradation ...  I subbed with new JJ multicaps with no measurable improvement, so I kept the originals.

I did replace all the bias electrolytics and all the coupling caps in the cathodyne. A lot of the old carbon comp resistors were all over the place, so replaced them as required too.
 

Attachments

  • Encel EL84 Integrated Stereo Amp Japan 1967 pic1.jpg
    Encel EL84 Integrated Stereo Amp Japan 1967 pic1.jpg
    303.9 KB · Views: 23
Here's a top pic ..
 

Attachments

  • Encel EL84 Integrated Stereo Amp Japan 1967 pic2.jpg
    Encel EL84 Integrated Stereo Amp Japan 1967 pic2.jpg
    309 KB · Views: 19
and front ..

The various mono and stereo modes are interesting to play with ...  great for comparing Beatles mono and stereo mixes  ;D
 

Attachments

  • Encel EL84 Integrated Stereo Amp Japan 1967 pic3.jpg
    Encel EL84 Integrated Stereo Amp Japan 1967 pic3.jpg
    249.2 KB · Views: 22
> Those crazy 60s Japanese hifi tube designer gurus for low cost solutions!

Fisher did the same many years before for a VERY expensive four-6V6 mono console.

You can find it all over, if you look.
 
That makes sense .. I had the impression the 'Encel' was based on a Fisher or a Scott integrated amp of that period.

I also think there was some connection with Sansui, like an oem done for the Australian market for a Melbourne based importer.

I will check out some more Fisher schematics.
 
The earlier Japanese import X1212 from Encel (dating from Sept 1964) didn't use that heater filament loading technique.

AWA used that technique in at least the PA872 and PA1005  PA amps - the earliest I can track those amps is the PA872 from circa Nov 1963 (first observed adverts in RTV&H).  AWA engineers would have had quick access to any new techniques of interest to them.

PRR, I couldn't easily identify the Fisher quad with heater filament loading - I looked through about a third of the schematics on http://www.fisherconsoles.com/
 
> identify the Fisher quad

Similar to the 1950 Coronet chassis R1:
http://www.fisherconsoles.com/r1%20console.html
http://www.fisherconsoles.com/r1%20console%20service%20manual%20page.html

Had a distinctly long quiet warm-up.

The one I saw had a Damping control. Ah, I see that in the 1955 Custom Electra I, which also has the quad-6V6 and preamp heaters for 6V6 bias.
http://www.fisherconsoles.com/custom%20electra%20I.html

Fisher Executive II Console (1958) is VERY different (styling, stereo), but if you trace it out the four EL84 in the power amp feed three 12AX7 in the control chassis.
http://www.fisherconsoles.com/executive%20II%20service%20manual%20page.html
 

Attachments

  • FisherR1-heater.gif
    FisherR1-heater.gif
    15.9 KB · Views: 22
Doh, don't know how I missed those - thanx. 

Ok, so the 1950 R1 has the cathode bias provided by three heaters in series, and a parallel 4k7 resistance (to provide fail-safe cathode voltage rise so as not to stress hk voltage rating), and a local 50uF bypass cap to constrain signal currents to the local output stage region and provide some possible distortion reduction (as sometimes found when bypassing the Williamson common cathode, and depending on the tube type used).

The 1955 K-15 appears to use some level of cathode winding feedback, and provides a fixed bias from what looks like a 2.2k to 1k divider off the bypassed DC heater supply - with the divider also providing fail-safe rise of hk voltage if an input stage valve is pulled, or gets a poor heater connection.

The 1958 Executive 2 is similar to K-15, but doesn't use the cathode winding feedback, and in addition elevates the ac powered heaters to the output stage cathode DC  supply derived from the preamp heaters.

A nice timeline and set of changes indeed!
 
Interesting ...


One thing I found a little unexpected in the Encel, was that the el84s, with their fixed grid bias + cathode resistance bias 'combo' ..  had no cathode bypass capacitors.

I added  1000uF to ground, at each of the cathode pins in the pp finals ....  made very little difference to the 3% power at the speaker terminals.

Usually when I bypass the cathodes, as I did in my 6v6 pp amp, I get a big increase in gain at the finals resulting in quite a bit more power to the speaker terminals. 

In the 6V6 with full cathode bias, the increase in power was substantial . .. like 60% or so. 

In the encel el84, with the combo bias,  it was more like 5%.

Again, I left it as it was, figuring it wasn't worth bothering with. I suppose that one needs full cathode bias to make bypassing effective.
 

Latest posts

Back
Top