Need some help conceptualizing switch for PS...

[Dr. Toobz]

I've started building my sealed "FrankenQuickie" which will end up being a small pcb for each channel, loaded by a 150H choke, which will be coupled to a SS output buffer/op-amp (via 3.3uF film caps) able to handle anything from 32 to 600 ohms. (I can't remember the chip # off the top of my head, but it's a common op-amp). I'm eschewing transformers and going for a hybrid setup this time, to get ideal performance with multiple cans, yet the same warm, choke-loaded sound I like from the 3S4 tubes (this time, as a voltage gain stage vs. outputs).

Here's the issue - how the heck would I switch on 4 different power supplies with the single push-button (on/off) switch on the pre-made chassis I'm using? There's the B+, two floating filament supplies, and then an extra supply that's basically just taken from one of the 9V batteries, going into a voltage regulator chip, to give the op-amp the 5VDC it wants. Replacing the push-button switch is not an option. Could I use some sort of relay on the power switch to get all four circuits to connect at once? Do they even make relays with so many connections? Otherwise, I can't envision how this would work. The filament supplies need to float, so it's not like I can take the B+ and drop it for the regulator, and then again for the heaters. That would be too easy :-)

I'm sticking to battery power for now, as I've yet to come up with a good way to get 4 supplies (36V, 9V, 1.5V, 1.5V) from AC. So, it's just getting one switch to do four things that I'm having a hard time picturing. Expert opinion needed!

EDIT: let me clarify my idea further - I can use the 9V going to the SS circuit to go through the power switch and a relay. I'd like to be able to use the 9V flowing through the relay (by turning on the power button) to switch on the other three supplies. Would I have to chain a few relays to do this?

[JC]

I don't imagine that you would have too much difficulty finding a relay with three or four poles that you can switch, but you will need one that latches in one position or the other, rather than consuming current the whole time it is "On".

Failing that, a "Push-On, Push-Off" type pushbutton switch with three or four poles would do the trick.  Look for something designated "3PST" or "3PDT" for three-pole, "4Pxx" for four-pole (in case you want to switch something else).

[Dr. Toobz]

Failing that, a "Push-On, Push-Off" type pushbutton switch with three or four poles would do the trick.  Look for something designated "3PST" or "3PDT" for three-pole, "4Pxx" for four-pole (in case you want to switch something else).

That's actually what I had originally intended to do, but the chassis I have coming has a built-in, SPST push-button power switch surrounded by an LED indicator "ring" - and doesn't look to be replaceable without destroying the case. So, I may be stuck having to use relays.

If the heater supplies didn't float (i.e., if I weren't using DHP's) I could have used the SPST switch to interrupt the ground right after the batteries, thereby shutting off the return flow of current. I could do that with the 9V batteries (both the B+ and SS supply from the first battery), but the filament supplies would require a separate switch. Unless, that is, I could un-float the power supplies and bias the tubes some other way. But I'm not real familiar with DHT's and this whole notion of the cathode bias and filament voltage coming as one package is difficult for me to get around. I'm still not 100% sure of why one cannot just supply 1.5V to the cathode, regardless of where it comes from (be it its own battery or following a dropping resistor from the main B+), so long as it's 1.5V above ground reference. I can see why these can't be wired in series, though, as the filament is the cathode for each separate tube. I'd think something like a 270 ohm resistor could easily drop the voltage from 5V (coming off the regulator chip) to 1.4V.

[JC]

Well, that is the issue with directly heated tubes, I think: You must have the incentive of a separate B+ supply to motivate the electrons emitted from the heated cathode to go where you want them to and do useful work for you.  Otherwise, if they are just going to be returning to the same supply that they are sourced from, they may as well just go to the + end of the filament, rather than make the longer and more difficult journey through space to get to the plate.

So, you will need to look for a low-current relay of 3 or 4 poles, and, as I said, it would be good to find one that latches in one direction the first time the coil is energized, then latches back the other way the next time the coil is energized. In other words, the relay needs to "toggle".

 On the plus side, none of the poles needs to switch very much current, so that should help with regard to size and power consumption.

Is it safe to assume that your existing push-button switch is "momentary", or is it "push-on, push-off"?

[Dr. Toobz]

Seems like all the relays at the local Rat Shack only "close" (or open, depending on how you look at it) when current is applied to the coil. Many of the ones I saw have a DCR of 160-300 ohms, and a current draw of 35-70mAh or higher. That's not good - the batteries will run down a lot quicker that way. What terminology do they use to denote relays that clamp one way or the other without continuously drawing current? I've looked in my Mouser catalog, but there's like a thousand different ones offered and I have no idea of what to look for.

I did all of my homemade PCB's last night - the voltage regulator board and a solid-state headphone op-amp on another (Burr-Brown) - and it sounds really good! Gain is set to unity (using 1M resistors) and that gave me a chance to rig up a temporary tube front end, which took some of the edge off the sound (the electrolytics in the signal path are probably the culprit). Better coupling caps - like 100uF film ones - at the output buffer, and the 3S4 front end will make this hybrid amp sound really great. Save for OTL amps, I'm starting to become a believer in this hyrbid topology for headphones vs. transformers, which I find to be finicky and a compromise at best. OTL amps are big and use more power than I want in this particular application.

If I can't make a relay (or even transistorized) switch to allow this setup to happen, I may as well ditch the tubes and just make a SS headphone amp with the Burr-Brown chip. It sounds great already, and would only require one power supply, anything from 6.8V-20VDC going into the regulator chip.

[Dr. Toobz]

Is it safe to assume that your existing push-button switch is "momentary", or is it "push-on, push-off"?

It's push-on/push-off.

[JC]

Do a search on "latching relay".  "Bistable" or "impulse" are also terms I have heard/seen.  "Toggle" or "toggling" might also work.

It is possible to make a standard relay latch, if you have an extra set of contacts on it, but that scheme requires energy to keep the relay in one position, which is what you would want to avoid with batteries.

Perhaps someone with more knowledge of the circuit can suggest another approach.

[Grainger49]

The nomenclature for the relay contacts is what the contact is when the relay is on the shelf, i.e. not in a circuit/picked (that means turned on).  What you describe will be hard to find.  A search of Mouser for industrial relays (14,396 found) yielded 309 9V relays.  The lowest power usage was 4mW.  This was one relay, SPDT, so you would need one for each supply.

[JC]

If I'm doing the math correctly, that would be something like .45 miliAmps, right?  That's not so bad.

[Grainger49]

That is the power required to hold the relay, inrush to pick it would be higher but transient.

[2wo]

How about something like an SCR or TRIAC? They will latch on with a single pulse…John

[Paul Joppa]

How about something like an SCR or TRIAC? They will latch on with a single pulse…John

N-channel power MOSFET with RDS(on) less than 1 ohm will easily take the 100mA of a 3S4 filament. Lower R is better of course, but losing even 0.1 volt still leaves a large majority of the battery life. Switch the gate to +9v, maybe 1 meg to ground so it shuts off when the +9v is disconnected. Ought to work!

[Paul Joppa]

So I looked for suitable MOSFETs at Mouser. There are hundreds; the cheapest one on the first page was the IRL630A  ($0.95) and it looks perfectly suitable.

[JC]

This certainly appears to me to be the most elegant solution, as well as being very cost-effective!

Are the grounds on all three power sources in the Quickie already tied together?

[Paul Joppa]

No, the three supplies (two filament and one "high" voltage) all have independent grounds. But the filaments float at a volt or two above the signal ground, so +9v on the FET gate will drive it to low resistance and allow the filament current to flow; restoring the gate to signal ground voltage will shut off the filament current.