Questions about Tube Regulation

[NC2W]

I'm an EE by day, but most of that time is spent wiring up pumps, fans & motors. Any help you are willing to lend is appreciated, but my understanding of vacuum tube tech is limited.

I'm finishing off winter, by building a DC supply for an AN/ARC-5 transmitter, to go along with my 2B.  I've looked around the internet, and found that the radio needs the following voltages:

530V@150 mADC, 260V @ 15 mADC, 200V @20 mADC, -50V@3.5mADC (bias), and some filament voltage (24Vdc).  I think I have the PA plate & screen voltages covered.  My question is really about the Oscillator Plate Voltage:

1.  Is a simple VR regulator tube setup, supplying 198V @20 mADC good enough to keep the oscillator clean? (my assumption is that it will be turned on and off, and the PA cathode keyed with a MOSFET). 

2.  I see here, from looking at several past conversations, people building modern semiconductor HV regulators.  If I use that approach, will I notice a marked improvement in signal quality (lack of chirp on CW, or AM audio (assuming l build a modulator later)?

Any help would be appreciated.
EWT - NC2W

[w4bfs]

hi Eric and I am also a EE (1980 TTU) and can understand how it gets confusing .... regulating the Osc supply certainly won't hurt things but I am puzzled about 200V spec as VR tubes tend to max out at 150V .... so I'll assume the 200V is the input to yer regulator ... the handbooks (old) laid this out but here is a quick app .... a OA2 is a 7 pin mini vr tube that needs 5 to 35 mA to operate ... so the Ohms law circuit is a series resistor to the vr tube which is parallel by the load .... assuming it needs 20mA at 200V (a 10 K Ohm load) then start the process with the assumption that it will also be a 10K load at 150V (15 mA)  this plus the the vr tube current requirement will then be 15 + 20 + 35mA.... The 20 ma is the max current of the vr tube minus expected load current.... now its simple as the series resistor  is set by R = 50/.035 = 1.5k @ 1.8W so use a 3W resistor

now if you want 200V regulated you will need 2 reg tubes in series with a 105V (OB2) in series with a 85V (??) vr and not sure if that might be a 5651 or sumphin like that .... I would probably to a FET at that point

hope this helps  73.  John

[NC2W]

I planned to use an OB2 & OB3, wired in series to a 265V supply, with appropriate current limiting resistor.

[WBear2GCR]

Close enough, good enough - if it is not, then change it!

[w4bfs]

ok seems reasonable .... the series resistor in this case drops around 100V and with previous current discussed would indicate a 3k resistor 5W ...

[WU2D]

I have used gas tubes to regulate the oscillator on ARC-5s, but you may have to double up to get enough current. It's kind of on the edge. To do this use separate feed resistors and diode OR the outputs. Later, I ended up doing solid state regulation with very simple zener and pass transistor follower setups. Bipolars out of old ATX power supplies work fine.

Separately regulating the screens of the 1625s at a higher voltage than the oscillator does pay off in terms of output power and chirp for CW. But for AM, just modulate the final conventional with a screen resistor. FMing is not bad if you have enough grid current.

[NC2W]

WU2D:

 The circuit I had envisioned was a tad different, but essentially the same.  I was using a dual winding 230VAC secondary (0.56A@460VAC), wired in series to develop 460VAC, at 0.5A, With a FWB rectifier, and a CLC filter network, I think I get (at least modeling using PSUD software) about 570Vdc, with the CT achieving 285Vdc.  I guess 285Vdc is appropriate for the 15 mADC going to the screens.  Seems like a lot of current for the screens, but that's what internet sources report.

I was planning to peel 20 mADC @ 285Vdc of Screen Voltage to power the oscillator, and allow it to run all the while it was on.  Ignoring the filament voltages, I had a total of (37.5mADC + 15 mADC) + 150 mADC of Plate Voltage = 210 mADC of the available 560 mADC of the secondary.  

I was using 1.5H of choke, and 55uF in the input, and 110uF in the output, split across both sides of the centertap.  The crude photograph shows the picture below:

The idea from this came from Peter Millett's website, based on a circuit he used for an 807 based AF amplifier.  I scaled the voltages up, using someone else's transformer and some software called PSUD.  Seems like large inductors are hard to find new, nowadays.  

I  like the 220V Regulator.  A lot simpler than some of Millett's designs, and appears scalable. 

[KA3EKH]

Some things are what they are, the ARC-5 transmitter was a fast-simple design to provide aircraft to aircraft communications in AM.
Never used one in CW mode myself but would assume that unless your willing to do a lot of work you are not going to get a clean “No Chirp” signal but so what? That’s part of the charm in using those old radios. They have a unique sounding chirp on the air.
OMRN (Old Military Radio Net) is on Saturday mornings on 3885 and you would be surprised the unique sounds a lot of those radios make.
A lot of the hard-core collectors are into running the radios with the dynamotors and that often results in strange keying noises and interesting modulation artifacts.
Don’t over think these things and get them on the air first and then you can go back and make improvements. The original design derived the oscillator voltage from the plate voltage so can imagine there was lots of variation of carrier frequency with modulation but in those days of wide band receivers and the like so what.

[N8ETQ]

Yo'


   Just String em up in series.  Thats fine.
Thunderbolt used a slew of em in series for
the 700 V screen supply.

/Dan

[NC2W]

Thanks for all of the replies.  They are helpful.  On one hand, I wanted to build a supply which would provide the transmitter a nice source, assuming that would 'clean it up'.  I have tinkered with using an HP-23A, but it seemed like 820Vdc on the plates was a tad too high. 

From looking at the schematics of these rigs, (and from looking at 1936 / 1955 / 1969 Handbooks) most oscillator B+ are sourced from the single anode supply, through a dropping resistor.  I thought (and probably it's not that sophisticated a problem) that modelling would tease out some behavior that I didn't already know from looking at circuits.  If we get this snow storm passed through, and MOUSER delivers my parts, I can work on the prototype circuit.  Then on into a case.

Again, thanks for the help

NC2W

[w7fox]

I have run a T19/ARC-5 on 80 meters for years.  It attracts lots of attention and is fun to use on the Classic Exchange contest.  I use a pair of 0C3s to run the oscillator at 210 volts.  However, it isn't close to being "clean".  There isn't any isolation between the oscillator and amplifier, and depending on how the final is tuned, it can put out an impressive chirp.  The oscillator runs continuously during transmit, and the final is cathode keyed.  I'm using the general instructions from the 1964 ARRL handbook.  I have read that it is even harder to get acceptable keying on 40 meters, but I haven't tried it.

http://www.radiomanual.info/schemi/Surplus_Handbooks/Getting_your_ARC-5_transmitter_running_-_Dave_Stinson.pdf
This article is the best I've seen for these transmitters.

[NC2W]

The power supply I am building is trying to satisfy the voltages called out in that paper.

[KK4RF]

I have a BC-459, which I tuned down to cover 80 meters. I made a power supply using WU2D's suggested solid state voltage regulator and it is virtually chirp-free even with keying of the oscillator and amplifier stages (separately using a fast DPDT relay.) Using regulated voltage on the oscillator plate and the amplifier screens. I also have a T-19 which I run with a HP-23A. I regulated the 250 volt supply of the HP-23 for the oscillator plate and amp screens, again using solid state regulators from WU2D's pages. I get a little chirp keying the oscillator, but if I let the oscillator run continuously during transmit periods, I can't appreciate any chirp  and I get good on-air reports.

I have a Command set that covers 40 meters and I have never been able to tame the chirp with this set. I may go back to it one of these days and work on it some more. Good luck---Marty, KK4RF

[W2JBL]

   I have two stations using the ARC-5 platform. One base, and one mobile. Both are plate modulated. I have tried regulated and non regulated voltage on the oscillator and never saw enough difference to make it worthwhile. My rigs run unregulated. As to voltage if you run 200 or so volts on the oscillator you don't get enough grid drive to the PA to make decent AM. Same for the PA screen. 300 volts on both works much better. Yes it's above the recomended voltage for the 1626 but in 10+ years of operation never lost a 1626. 500 volts is a bit wimpy for the PA plate. I run the full 800+ volts out of a Heathkit HP-23. There is also no need for a negative bias voltage. The HP-23 makes enough to power an ARC-5 to 100 watts out as well as supply a complete modulator with triode connected 6146 tubes in class B.

    No need for a mosfet to key the cathode. ARC put a nice relay in to do that. It will work on 12 volts too. It keys the oscillator plate and PA screen and follows fast CW keying no problem. If you try to use CW with the oscillator running full time you will have a very healthy backwave at your output due to the PA not being fully neutralized.
   

[NC2W]

When you say that no bias is required, were you simply leaving the bias pin disconnected, or ground it?