Filament Power, for larger builds

[aurt]

Why is that? In the distant past there were lots of battery radios that used directly heated tubes.

Cheers

Ian

I suppose you can, but you run into some interesting bias problems when one side of your cathode is x number of volts higher than the other. It can also cause uneven wear/stripping. I think there are reasons they stopped doing it in the distant past.

 

[RoadrunnerOZ]

As for distance of supplies from the console, how far is too far? I have 600V 16 AWG wiring i use for HT leads which seems to work great in other projects, and then will have the just mentioned 14 AWG wiring (shielded one end) for Meanwell. Is 6-8 feet too long for either of those runs?

6 - 8 feet is not a problem - I’ve run 30 - 40 feet using power multicore wiring consoles with the supplies in another room.
The MSP’s can tolerate current overload and recover automatically like the LRS just without hiccup mode - the use of an over-rated supply should preclude this but for normal every day use I’d be using the LRS anyway - had great success with these - same as seen in the preceding photo - the supplies I chose after doing a clamp meter test to read the turn-on surge and running current of the Neve - 350W for 15 and 24 and 500W for the +16 - the fast blow fuses in the original power supplies were a good indicator of what current rating to use anyway. Been running every day for 4 years and no issues. Drawer temperature 27°
I also have a couple of Meanwell multi-voltage supplies I use for quick testing, all sorts of mistreatment and they have been going for 5 years so far.
Before buying a supply you can always hook up a car battery and get a read on the start and run current for all filaments together.
These LRS supplies are cheap enough to be a painless replacement if they blow. I think $22 in the USA for what you need

 

[RoadrunnerOZ]

I suppose you can, but you run into some interesting bias problems when one side of your cathode is x number of volts higher than the other. It can also cause uneven wear/stripping. I think there are reasons they stopped doing it in the distant past.

Indirectly heated cathode doesn’t care about the filament polarity - if there was enough filament to cathode leakage to cause a problem with dc then with ac you’d have massive hum problems.

 

[Rusan]

You can’t run directly heated tubes that way, but otherwise it’s a fine idea with definite benefits. I did get one DHT amp in once that had the 300B heaters run from a 40kHz oscillator and SS power amp through an interesting handmade toroidal transformer to give them a center tap ground. I guess if you really really don’t want any 60Hz in there, that’s one approach.

Why is that? In the distant past there were lots of battery radios that used directly heated tubes.

Cheers

Ian

I suppose you can, but you run into some interesting bias problems when one side of your cathode is x number of volts higher than the other. It can also cause uneven wear/stripping. I think there are reasons they stopped doing it in the distant past.

The voltage gradient across the cathode also carries the caveat of reduced emission on the low voltage end. The Japanese DIYers who pioneered the single ended DHT amplifier renaissance back in the 80s found that directly heated power triodes sounded flat and dull when heated with DC, with the ubiquitous 300B being notorious for it. I've heard several DHT amps with AC filaments and even built a couple myself, but none with DC, so I can't comment from personal experience. But, countless reviews (including from some trusted friends) of DC-fired DHT amps have all agreed with this.

In the early 2000s, some of the more serious DIYers and then a few manufacturers in the U.S. started powering their filamentary cathodes with ultrasonic AC oscillators, like Aurt had mentioned. One particularly interesting circuit was based on a UC3872 push-pull resonant lamp ballast controller, which drove a small transformer with a capacitor placed across its primary to tank at about 100kHz. The idea was that the 100kHz "hum" couldn't couple across the low-bandwidth single ended output transformers. Therefore, no chance of ultrasonic voice coil heating in the crossover-less, single driver speakers commonly used with flea-power amps.

 

[aurt]

Indirectly heated cathode doesn’t care about the filament polarity - if there was enough filament to cathode leakage to cause a problem with dc then with ac you’d have massive hum problems.

Indirect is fine. I was speaking of directly heated tubes.

 

[ruffrecords]

Ah, thanks for spotting that. Was otherwise going to go with the medical one.

As for distance of supplies from the console, how far is too far? I have 600V 16 AWG wiring i use for HT leads which seems to work great in other projects, and then will have the just mentioned 14 AWG wiring (shielded one end) for Meanwell. Is 6-8 feet too long for either of those runs?

I regularly have a 6 foot run from an external power supply to a tube mixer

Cheers

Ian

 

[emrr]

Then there’s this school of thought with DHT’s and DC. Several later DIY projects along those lines.

http://www.tentlabs.com/Products/Tubeamp/Tubefilament/index.html

 

[aurt]

After a bit of digging, I’ll retract my earlier comment on DC DHTs. Seems it’s less of an issue than I had thought unless you’re running them very hard. I still like the ultrasonic method though.

 

[systemtruck]

In the context of having both a HT/rectifier supply, and a DC 12.6V supply… i suppose all chassis earth should go back to the HT/rectifier supply earth which then goes to the wall outlet…. and I suppose the Meanwell own supply’s earth input from wall should be strapped to the DC 0V of its output?

That would create two isolated earth situation roaming around the console though, making contact with different pins of the same tubes. So maybe it’s best to not strap the DC 0V to the Meanwell earth and just let the 12V/0V pair float freely throughout the board?

I’m not sure how the Meanwell handles ground internally. Maybe its 0V output is already connected to its earth from wall.

 

[RoadrunnerOZ]

In the context of having both a HT/rectifier supply, and a DC 12.6V supply… i suppose all chassis earth should go back to the HT/rectifier supply earth which then goes to the wall outlet…. and I suppose the Meanwell own supply’s earth input from wall should be strapped to the DC 0V of its output?

That would create two isolated earth situation roaming around the console though, making contact with different pins of the same tubes. So maybe it’s best to not strap the DC 0V to the Meanwell earth and just let the 12V/0V pair float freely throughout the board?

I’m not sure how the Meanwell handles ground internally. Maybe its 0V output is already connected to its earth from wall.

No - the earth is separate from DC 0V in the Meanwell supplies, the Meanwell supply chassis should be grounded at its own end, but 0V should be free of the Meanwell mains earth unless the supply is not remoted - if any grounding of the filament supply is warranted then ground the 0V line at the amp chassis wall earth star point. I have done this with console remote supplies - the supplies have their own chassis earth but no DC line 0V earth connection at the supply end which could cause a hum loop. You could add an electrolytic as extra filtering at the amp end as well as 0.1uF caps at each tube filament.
Edit: read mixer instead of “amp” chassis

 

[ruffrecords]

In the context of having both a HT/rectifier supply, and a DC 12.6V supply… i suppose all chassis earth should go back to the HT/rectifier supply earth which then goes to the wall outlet…. and I suppose the Meanwell own supply’s earth input from wall should be strapped to the DC 0V of its output?

That would create two isolated earth situation roaming around the console though, making contact with different pins of the same tubes. So maybe it’s best to not strap the DC 0V to the Meanwell earth and just let the 12V/0V pair float freely throughout the board?

I’m not sure how the Meanwell handles ground internally. Maybe its 0V output is already connected to its earth from wall.

MeanWell supplies tend to connect frame ground to the input EMI filter via a capacitor and also to the output 0V via another capacitor. This helps CMRR.

For general mixer grounding see the attached Grounding101

Cheers

Ian

 

[aurt]

Ian,

This is why the AES recommends connecting pin 1 of XLRs directly to chassis(and not 0V) at the connector.

I remember the Pin 1 Problem paper from, the 80s or 90s, fairly well. I have a friend who builds guitar amps that still elbows me a bit for drilling common impedance coupling into him.

Edit: '94 AES paper by Niel Muncy.

 

[RoadrunnerOZ]

Lot of gear got modified to fix the pin 1 issue back then. Some studios were cabled for balanced gear so the patchbays and equipment runs had the screen grounded at the input of any device only and screen ground lifted at the output end of any chain (except for mic inputs).

I’ve always treated patchbays as an extension to the console and left the grounds separately wired at each patch point, not common bus grounded at the patchbay which can cause problems, but some outboard gear would have the output grounds lifted in the hard wiring if needed.
(There were also schools of thought the opposite way - lifting input grounds and leaving output grounds connected - not sure how well this would work)

 

[ccaudle]

There were also schools of thought the opposite way - lifting input grounds and leaving output grounds connected - not sure how well this would work

MisterCMRR published a paper showing why connecting at the output and lifting at the input was the best way to make a shield-at-one-end connection.
I think it was in the same June 1994 special issue of the JAES that popularized the term "pin 1 problem."
Bill also put a page in the presentation he has linked here a few times before:

 

[ccaudle]

By the way, just so no one is confused, this part of the picture you clipped in earlier is exactly the wrong way to connect a shield. It demonstrates how to create the "pin 1 problem" which has been warned about extensively over the past 30 years.

 

[RoadrunnerOZ]

By the way, just so no one is confused, this part of the picture you clipped in earlier is exactly the wrong way to connect a shield. It demonstrates how to create the "pin 1 problem" which has been warned about extensively over the past 30 years.


View attachment 141739

That diagram is from an article by a guy called Ben Duncan - I imagine it just shows a generic bal input with supply ground to pin 1.
When wiring wall plates for external gear connection to the patchbays I would always have grounds carried through and for hard wired gear the same. If any issues arose with hum/noise, short test cables with shield pin 1 lifts were used to determine whether to lift shield grounds at input or output (if that was the problem).

 

[CJ]

If you use AC for heaters you have the ability to put in a hum canceling pot which can be used to cancel out noise if working in a high density 60 hz zone.

 

[Markku01]

Hello,
I’m taking steps towards a build that will power up around 20 or so tubes in one unit.
It’s all 12.6V heaters.
I have a toroidal “tube transformer” that offers 4A on its 12.6 VCT secondary and I am considering using 3.6A of that 4A.
The HT secondary of same transformer offers 550mA, but I’ll only ever use around 150-200mA of that.

Would be it smart to get a separate 12.6V supply? Maybe something like a Hammond that provides 10A?

I’m sure I can “get function” with using the onboard 4A as mentioned, but will that constant near capacity current draw from one secondary be a real slog on the other secondary that’s providing the HT?

And will it also reduce the life of the transformer?

I might want to run things as cool as possible. So maybe a 10A filament supply is the better choice.

Yes, 10 amps will run a lot cooler.

 

[aurt]

Lot of gear got modified to fix the pin 1 issue back then. Some studios were cabled for balanced gear so the patchbays and equipment runs had the screen grounded at the input of any device only and screen ground lifted at the output end of any chain (except for mic inputs).
View attachment 141724
I’ve always treated patchbays as an extension to the console and left the grounds separately wired at each patch point, not common bus grounded at the patchbay which can cause problems, but some outboard gear would have the output grounds lifted in the hard wiring if needed.
(There were also schools of thought the opposite way - lifting input grounds and leaving output grounds connected - not sure how well this would work)

Not to stoke the threadjack too badly (too late), but the Pin 1 Problem is something extant inside the gear and has little to do with the interconnects. While it might be somewhat mitigated by single ending the shields by lifting them at the gear end, the problem itself has more to do with it being cheaper to manufacture gear with PCB mount XLRs that connect shield to the audio ground rather than to chassis ground. The gear itself needs to be modded internally to fix it. You need to cut the trace and run a wire to the chassis. Not sure if that's what you're talking about.

Sadly, Muncy's '94 AES paper isn't easily findable free online. It is exceptional.

There are different schools as to best practices for what to do with the shields. Up to and including landing one end and connecting the other through a small cap for RF. Always sounded like a pain to me.

A place I worked 25 years ago was set up with male end lifted, bussed shields, if landed, at the bays. Mic lines excepted. Still had noise issues. I redid quite a bit of stuff there this last year, some fairly guerilla, two rooms very conventional. Landed everything, bussed at the bays, and have had virtually no problems (one issue with the monitors on a TSM aside). Don't ask me what changed. The place was built in the 70s, and redone several times since. Oddly, it seems to work best like this. Just goes to show that "ground loops" aren't always ground loops. Nowadays, I tend to land, and lift if I have problems. I find relatively few.

But again, a whole different animal to the pin 1 problem.

That diagram is from an article by a guy called Ben Duncan - I imagine it just shows a generic bal input with supply ground to pin 1.

Which is, as was mentioned, the definition of the pin 1 problem. Ian's linked paper touches on the issue well.

 

[RoadrunnerOZ]

Yeah - as I mentioned before a lot of studios had output ground lift mentality back in the 70’s and 80’s, some gear was actually made with ground-lift switches, but there were also issues from unbalanced aux sends and returns going to/from balanced and unbalanced gear, installation wiring and patchbay wiring methodology Also was the problem with input ground lifted cabling creating noise if not connected to anything except an input - there was the problem of RF isolation vs hum loop correction. It seemed like there was no rule of thumb you just dealt with each piece of gear in a certain wiring environment and modded as necessary. Some issues were fixable with via cabling, others required modifications to the gear - certain series chains would have problems like desk>compressor>eq>desk.