Tube Mic PSU Heater Voltage Mod...

Hi all,

So I want to mod the attached PSU for ~6.0V heater voltage. I've got a few other questions as well if anybody could chime in. I will be building this from scratch into Sony camera power supply cases for my M50-like mics using the Raytheon 5703WB tube.

Sorry for the large pic, but you can't read it otherwise.



Question components highlighted in red.

1. Gr1 - Diode - Looks like they used two in series here; presumably to up the voltage? So I could use just one 250V rated diode?
2. Gr2 - Bridge Rectifier - Just use 4x IN4007 or something?
3. Gr3 - Zener Diode - Any clue as to the value of this diode? It's used to regulate voltage right? So I want 120V?
4. R5 - ? - Is this a 50k linear trim pot? Sorry to sound stupid but want to make sure.
5. R12 -- Since my mic is fixed-pattern, can I just omit R12, R15, C8, and the pin 6 connection? Do I need to replace R12 with a fixed resistor?
6. Gr4 - Zener Diode - use 6.0V zener for desired 6.0V output? Can tweak with R9, but do I need to change anything else to handle a 6.0V 200mA load from the designed 4V 100mA load?
7. R14 - ? - Another trim pot here? Note the symbol difference with R5... Or wait, this isn't really a pot -- only 2 leads?!
8. TRS1 - AD130 regulator IC - I can still get this part. Any reason I should get something else?
9. Power Transformer: Where in the heck can I find a dual secondary xfo with 210V and 20V? Been looking at the smaller toroidal pieces from the likes of Amveco and not finding anything. I guess I could use 2 separate pieces...

For convenience here is the modified mic schematic to go along with this project. The original mic heater with AC701 tube pulled 100mA at 4V. Sorry for the long post but I tried to organize this the best way I knew how.



Thanks for any ideas guys!

Hi Riggler,

just curious, why do you want to build a PSU with these old schematics? It won't make any difference in sound if you modernize a bit, and both parts and schematics are easily available.

Martin

Hello Martin,

Well I had thought of doing that, but Gus had illustrated that there is some very good engineering in these older designs. In this mic application, the heater and cathode are tied, so a very clean filament voltage is necessary.

I'm all ears as to a more modern topology, but I need very clean power. I wish I had the engineering experience to better know the intricacies of how design affects quality. I mean, I get the basics of transformers and bridge rectifiers, but I don't know which regulator is better than what, etc...

I just read "Power Supply Design Basics" by P. Antoniazzi, and the author states that "Zener-based Regulation can supply load currents up to about 100mA. So maybe this design won't work so well for me since I'm pulling 200mA heater...

Thanks for any info! I was looking at Jakob's G7 mic PSU but someone had said it may not be quiet enough for my application...
What about Max's? Would that be good? (Then I'd have to rework it for the proper voltages...)

1. Use a bridge rectifier for the B+. Its easier to filter. This means the B+ will be slightly higher so use a transformer with about 190V output (B+ = 1.414 x tranny RMS).

2. Yes, 1N4007 (or 1N4004 or 1N4005) okay.

3. Its 150V. You can use two that sum to 150V.

4. Yes, linear.

5. Yes, you can eliminate all the pattern components. You might want to replace the pot with a resistor, between 100K and 220K just to help discharge the supply when off and prevent the roadies from launching themselves across the room when they disconnect the cable from the mike and lick it (I don't know, its a roadie thing).

6. I think its around 2.5V if the voltages shown on the schematic are correct. See later comments.

7. Its a pot. They just connect the wiper to one of the ends.

8. Don't bother. Use a modern transistor. The filament current for a 5703 is very low.

9. In the true DIY style, I wind my own. I start with a Toroid Corp. core kit, lots of wire, a couple of popsicle stick bobbins, and I watch reruns of "Deal or No Deal" (I like that because I know already how it's going to end).

The B+ section of this circuit is fine except for the changes noted above. I would use a packaged voltage regulator, like an LM317, for the filament section with a huge filter before it (>10,000 ufd) split into three sections as shown, with resistors between. My current mikes use 20,000.

I did not post copy the circuit 100%.  

riggler said:

In this mic application, the heater and cathode are tied, so a very clean filament voltage is necessary.

Click to expand...

You are absolutely right about this, and while you always have the option of using a later version of the Neumann circuit that has the heater separated, neither a simple zener or LM317 regulator will be low enough in noise. Not talking about hum here, hiss from the regulator is more likely the problem.

What might work for you is a simple solution, start with a normal regulator circuit (such as the G7 PSU) and follow it with some RC stages as needed to bring the regulator noise down. This is low voltage, so you can have lots of capacitance, up to Farads in the form of those memory backup supercaps...

Just set the LM317 or whatever to supply enough voltage so that you end up with the correct heater voltage at the tube, after all the voltage drops over the R's. Since you know the heater current you can calculate the voltage drop.

Martin

Or one could read a book or two about power supply design and build a good supply.  I don't like the G7 power supply but that is me if one searches you will find problems with low B+ IIRC I think it is due to VA ratings not matching the ones used in the first G7 build.  That said the G7 was on the web years ago and gave people a nice web page to build a microphone.

Years ago Brad posted a very good link to a page for low noise power supply designs the person listened to the supply rails and checked the waveforms IIRC .  

Look at what is going on with the first posts linked schematic fil supply
Series or shunt?
and other things

I have used super caps in a supply.

Don't forget diode switching noise.

riggler said:

7. R14 - ? - Another trim pot here? Note the symbol difference with R5... Or wait, this isn't really a pot -- only 2 leads?!

Click to expand...

burdij said:

7. Its a pot. They just connect the wiper to one of the ends.

Click to expand...

Hi,

it's an old symbol for NTC, IMO. It's used here probably to compensate the output voltage drift
caused by the transistor's temperature change.

Regards,
Milan

these are some bullet proof heater circuits from a thread a while back...

i can tell they are not where you are headed, just a FIY thang, you know...

for prosperities sake, or whatever those edjucated people say, all fancy and sh*t....


http://www.groupdiy.com/index.php?topic=728.0

> "Zener-based Regulation can supply load currents up to about 100mA.

Oh, bosh! My BSA moto-cycle regulated its 50 Watt (12.6V 4A) dynamo with a huge 50 Watt Zener shunted across it.

Like most brute-force Zener plans, heat came out all over. The Zener was bolted to the frame, actually on the oil tank. (The frame was both oil reserve and cooling.) The excess heat appeared in the dynamo windings, but they were in alloy on an engine dumping 10,000 Watts heat from every fin and cover, so another 30W-60W went un-noticed.

You want 200mA to the load, and maybe 100mA nominal in the Zener to allow for raw-supply variation. When the load falls off, you have 6V 300mA in the Zener. It must be rated almost 2 Watts. That -is- awkward. You might find 5W Zeners.

You can boost a small Zener with a Power transistor. But that's still brute-force waste-heat. And there are better paths.

The NN48h supply is not afraid of waste heat. It gets 4.2V from (18.5V+4.9V)= 23.4V. At least 82% of your electric bill is in the box, NOT in the tube heater.

There's history behind that. In 1967 they could not get (or did not trust) LARGE capacitors. C2 R2 C4 R4 C5 form a heavy filter with 500uFd caps and lots of voltage drop (heat). We could do it different today. But for a one-off in an ample box, I don't see much point. I -do- like my many-cap filters.

Gr4, the Zener, obviously has (2.1V+4.9V)= 7V across it. R6 obviously has (8.5V+2.1V)= 10.6V across it. R6 is 120r, so the current is 10.6V/120= 0.0883A. Assume all of this flow to the Zener. The heat is 7V*0.0883A= 0.62 Watts.

A contemplation of voltage and current around R9 R10 confirms that most of R6's 0.088A flows in Gr4.

So Gr4 is 7.0V >0.6W.

R9 R10 also work with C7 to filter Zener hiss. It is a ~~2Hz low-pass. 20Hz flutter is reduced, 100/120hz ripple is reduced a lot, and >2KHz hiss is way-way down.

The whole heater supply seems "upside-down" to modern eyes. The transistor emitter is common. If Trs1 were an NPN, and all else polarity swapped, we'd common the other side. But into 1967, Ge PNPs were widely used in power applications. Those parts have faded, but an Si PNP is cheap, and avoids the brain-pain of correctly swapping all polarity on this plan.

Note that the scheme really makes regulated 5.1V. The drop to 4.2V is taken in R16 R18. This may be to control cold-start heater current, or may be the short-protection (you MUST assume mike cables can be shorted); it can't dump more than 5.1V/(1.5r+5.1r)= 0.77A short-term, so Gr1 could be a 1A part. (Long-term, R2 R4 limit current to 0.4A.)

K.I.S.S.

Use a 6V lantern battery. It will last a long time. It is very clean. There are only two wires to mis-connect. It blows-up gracefully, and a very-hot dead battery is easily diagnosed and "repaired" (replaced).

You have many other issues to work through, and this project won't be "finished" soon. Drop a dang battery on it and get some current flowing. Worry about wall-power when battery costs are really denting your studio profit.

CJ thanks for the link. The trick with the memory capacitors seems especially handy:



1. Those things are tiny so you can "upgrade" this little gadget into any mic heater supply.
2. The 5.5V versions are quite cheap.

Hey guys just got through reading all of the replies here and I can say I am officially overwhelmed.
I think I might just do the 6V lantern battery! My cases are pretty big and I can make a nice "home" for them, then I just need to worry about B+ which is not so critical.

If I use a lantern battery wouldn't I have to worry about the voltage dropping slowly over time? Or is it going to be inconsequentially minute? Regulate it?
When turning on, should the battery be connected to heater before or after B+ is switched on? Can I use a DPST switch to turn both on at the same time?

I'll draw up a scheme of what I plan to do for B+ and see if you guys see any issues with it.

THANKS!

This is the "Drawing Board" that is why I did not give answers.

This more of a lab thread IMO.  If you want to build something simple maybe do what Martin posted for the heater supply.


EDIT
http://web.archive.org/web/20080703174319/http://www.tnt-audio.com/clinica/regulators_noise1_e.html

Hi all,

Okay, so this is what I came up with. In one of the articles Gus suggested they were able to much reduce noise from a 317 by putting a bypass cap on the adjustment leg. So, I used the datasheet diagram for a "high-ripple-rejection" supply, added the bypass cap, and also added RC filtering ahead of the regulator. Since we're talking low currents here, I've read that I do not have to worry about adding a protection diode over the adjustment lead.

Of course this makes me ask more questions:

I sometimes see RC filters using non-polar caps, and then I see them using polar caps in PSU's like this. Why?
Am I doing too much filtering now? I don't think the voltage drop will be too bad, seeing as to how we should have around 30V coming off of the bridge rectifier.

Does this look good? I'm not too sure of my resistor values in the RC filters. I've done an online simulation showing that about 40 ohms, 3300uF, is good to down about 100Hz. Sound good?

Per the original mic PSU schematic, the negative leg of this supply gets together with the negative leg of the B+ supply at pin 7. Is everything still cool with doing that? I ask because I'll be using 2 separate transformers here. One for B+ and one for filament.

OK I'm up for any pointers. Thanks all!

First thing that comes to mind is the rule not to use tantalum caps in a power supply. When they fail, they go short circuit. Maybe new tantalums have not this problem, but I would still use standard film caps (mylar, polyprop, whatever) for safety reasons.

A possible problem is that measly 1uF cap at the output. Might not be enough to take out the last of the regulator noise so maybe use another RC (like those 40ohm+3300uF) stage after it.

Another thing to consider is that a cold filament has a very low resistance, so it seems a good idea to put a current limiting / dropping resistor in series.  The Neumann U64 supply (for the 7586 tube) uses a 33 Ohm 3 Watt resistor with a 10 Volt supply and a filter cap on the back side.
Bruno2000

Edit: Bad grammar.

Will change the tantalums to something out. But t's funny as the data sheet specifically calls for them!
As far as handling inrush, do I use a thermistor here, some other type of special resistor, or just a regular resistor??

You have to use RC filtering after the regulator. And before the regulator you only need CRC, anything extra is waste of components. Properly used LM317 is very good in ripple rejection, but noisy. I measured 15uV RMS white noise from a circuit like yours with 2200u cap in output. Way too much.

In general long CRCRCRC chains with identical R and C values don't make much sense IMHO.

You can make a soft start by using big cap (4700-10000) for the ADJ pin if you feel like. Just an idea.

riggler said:

Will change the tantalums to something out. But t's funny as the data sheet specifically calls for them!
As far as handling inrush, do I use a thermistor here, some other type of special resistor, or just a regular resistor??

Click to expand...

The Neumann PS uses a 33 Ohm WW 3 Watt with one 500uF on the back side, then out to the mic.  The 10 Volt regulator is a Zener.  I used a 317, with 2200uF caps, and it seems to work fine.  I wanted to keep the old engineering, just use better parts.
Best,
Bruno2000

Okay, so the noise of the 317 is broadband-type noise, right?
Wouldn't an RC filter after quell that (down to the corner frequency at least)?

Bruno, do you have a scheme or drawing of what you did? In your mic, is your cathode and heater tied? If so, how's the noise? After the 1uF cap, I do another RC using a 33Ohm wirewound 3W resistor and 500uF cap. I'm not adding a zener here -- you meant you replaced the stock design zener regulation with a 317. Right?

I think I'm going to get rid of one of the RC filters up front, or change it's value to something else to get me filtering down closer to DC...

Jonte, like the idea of the big cap for soft start. Just plop it there in series with the 220R resistor right? You have me worried about noise. Would the above mentioned post 317 RC filter be enough? Gotta start building now to play with this and buy and learn to use a scope...