Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: [1]   Go Down
  Print  
Author Topic: High Voltage question  (Read 3478 times)
N4UE
Member

Posts: 709




Ignore
« on: November 25, 2003, 02:36:35 PM »

Hi to all. I have been a ham for 40+ years, but don't have much experience with high power (tube) amps.
Years ago, I bought a monster 6 M amp from a fellow. I re-did the input circuit so the SS radios were happy. I replaced the two 3-400Zs with 3-500Zgs. My question concerns the drop in the HV. When in stand-by (with the tubes biased off), the HV is 4KV.
(the amp has a HUGE plate xformer and ganged Variacs to control the 220 VAC). When I Xmit a 100 W signal to the amp, the Plate Voltage drops to around 3 KV. At this point, the amp is producing just over 1000W output. Because of the Variacs and additional taps on the transformer, almost any voltage is possible. My handbook says the Plate Voltage on the 3-500 to be 4 KV, so that's where I run it.
I was just wondering if anybody out there could tell me what kind of a voltage drop they see in amps using these type of tubes.

The P/S in this old girl is very old. It is about 3.5' high, and 2' square. I suspect it weighs about 400 lbs. It has 2 of the biggest chokes I have ever seen to filter the B+. The HV transformer (pole pig?) is a monster and weighs over 100 lbs. The rectifier assy is made up of LOTS of individual diodes, in series (4 legs), all fitted with voltage equalizing resistors and ceramic caps. I have 4 of the "Silicon Alley" very high voltage (and current) modules. I have heat sinks to mount the modules on, etc. Still, the original builder (a retired machinist) did such a wonderful job building this amp, I am almost hesitant to change anything.

I would just like to see the B+ stay a little more constant.......

thanks!

ron

N4UE
Logged
WB2WIK
Member

Posts: 21754




Ignore
« Reply #1 on: November 25, 2003, 03:41:18 PM »

That is a rather huge swing in plate voltage, especially for such a monstrous power supply.  Then again, if the builder was a machinist and not an electrical engineer, he may have known more about cabinetry than about power supplies.

The *maximum* rated Ep for a 3-500Z is 4kV, but most amps run them at 3.0 to 3.5kV, where they can still develop full gain and output power within rated Ip limits.

My 3-500Z(G) amplifiers run ~3200Vdc key-up and ~2900Vdc key-down, for regulation of about 10% no-load to full-load.  That's typical, and what most 3-500Z amplifiers do.  If your amp is really running 4kV no-load and 3kV full-load, that's worse than 25% regulation and not very good at all.  Probably not hurting anything, although regulation this poor often indicates lack of adequate storage (filter capacitance), which can result in hum modulation of the amplified signal, especially on CW.

If the power supply has chokes as you say, the poor regulation may simply be cumulative IR losses from the resistance of all those windings.  I've never used a choke in a high voltage power supply in 35+ years, and don't intend to ever do so.  I just use lots of capacitance.  Chokes used with solid state rectifier power supplies are kind of silly and originally were used to limit surge current which was damaging to early silicon rectifiers.  That problem went away in about the 1970's, so I have no idea why anyone uses HV chokes for this application anymore.

If you can key-down (send a long DAAAAH on CW) with this brute and a receiving station can't hear any hum modulation on your signal, you're probably doing fine and really needn't change anything.

WB2WIK/6
Logged
N4UE
Member

Posts: 709




Ignore
« Reply #2 on: November 25, 2003, 06:01:25 PM »

Hi Steve. I appreciate your answer. I did a test earlier today, (being VERY careful around these type of voltages). I also thought about the IR losses in the chokes. Soooooo, I put a jumper across the chokes, one at a time. The result was unexpected.... The standby plate voltage (PV) dropped to 3 KV. Under full power, the PV dropped to about what it was  with the chokes in-line. I did not expect that result.

I have another question for you: When tuning up (into a dummy load, of course), the 'Load' cap produces a nice peak at about 1/2 capacity. Good. But the 'Tune' cap is so broad as to not really have a peak. It's a very nice Vacuum Variable (VV). When the amp is showing max output, I have run out of travel. Minimum capacity, (maximum spacing). Using my NIST calibrated cap bridge (and subtracting for stray cap. in the leads), the cap measures 10p to 50p. When I run out of travel, the cap is at maximum gap, that is around 10p. Here's the rub. The original builder (yes, he did a WONDERFUL job on the cabinet, it's a work of art), built a 'plate sandwich' to connect to the end of the VV. One end of this 'sandwich' is an alum. disc which connects to the VV, then there are 4 transmitting caps (in parallel) in the middle of the 'sandwich'. The other end is another alum. disc. This goes to the inductor, which I built and had silver plated. The measured capacity of this 'sandwich' is about 350p. Having this in series with a 10-50p VV obviously swamps it. When dipping it, it has a nice dip right where needed, but the 'tune' VV has little effect. This value (this is the first opportunity I have had to 'investigate' how the amp was made) seems way off.
My question is: what is a good ballpark value for a 'Tune' cap for a pair of 3-500s? I believe your answer may be in another thread discussing the conversion of an SB-220 to 6 Meters. I do have a good supply of transmitter quality variable caps.
This P/S has some type of voltage regulation lamps (??) in the HV section. They are about 12" long and look very much like a florescent lamp, except for the ends. There are a bunch in series-parallel.
Again, the builder was a machinist, not a EE. The cap bank for example, has lots of metal cans (probably with PCB!) with nice insulators on the terminals.
The amp does give consistantly good signal reports. When working you on MS (back in the 80's, when you were up 'north'), it made all the difference between no copy and residual S2........
I just want maximum efficiency. I know I could rip out all the old stuff and start over, but....... I do have some good test equipment, but this is my first opportunity to 'work' on this old girl in the 20 years I've owned it......

Lastly, as a paracitic suppressor, the builder had strips of copper flashing running from a choke to the tube anodes. Near the tubes, mounted to the strips, are resistors, obviously there to suppress parasittic osc.
This amp was a duplicate of a conversion Bill Orr did on a Swan amp. The suppressors described above were in the original article. I have it. In fact, I spoke will Bill about the design of this amp and it's efficiency, before he became a SK.......


thanks for your great expertise, as usual!

ron

N4UE
Logged
WB2WIK
Member

Posts: 21754




Ignore
« Reply #3 on: November 25, 2003, 08:06:38 PM »

Well...

The DC output voltage from a full-wave rectifier with a capacitor-input (or capacitor *only*) filter is 1.414 x the RMS value of the transformer secondary.  The DC output voltage from the same full-wave rectifier but with a choke-input filter is 0.9 x the RMS value of the transformer secondary, assuming you have enough bleeder current to keep the choke magnetized.  What you explained doesn't make much sense to me, either, but then maybe the circuit isn't so obvious.

Anyway, the TUNE cap for a pi-net on 6m using a pair of 3-400Zs or 3-500Zs should be very, very small: On the order of 3-4 pF.  Which is why I normally use a 2-30pF 5kV or 7.5kV vacuum variable.  It sounds like the builder attempted to reduce the minimum capacitance of the vacuum cap he had with the series "sandwich" cap as you describe, but maybe didn't go quite small enough.  If you can replace all that stuff with a single 2-30pF Jennings vacuum variable, keeping interconnecting inductance very small (short, broad conductors -- not braid, but copper strapping), the whole circuit might work better.

BTW, braid is not such a great conductor of RF, especially at higher frequencies, and should be avoided for interconnections at 50 MHz.  Also, if the builder of the amp was more of a mechanic than an engineer, he may not have realized the value of keeping the RF ground really, really close to the tubes and plate tank components.  You can't rely on going through an aluminum panel, down to an aluminum chassis, and then several inches along the chassis, for grounding -- at DC, it's zero Ohms, but at 50 MHz, that path can have a lot of reactance.

73,

Steve WB2WIK/6
Logged
WW1Z
Member

Posts: 35




Ignore
« Reply #4 on: November 26, 2003, 05:01:19 PM »

Steve's comment about the expected voltages sound right to me. So in a proper choke input supply shorting out the chokes should increase the voltage. If there is enough current through the chokes. Here it seems shorting out the chokes dereases the voltage. Sounds to me like the chokes and filter capacity may be resonate at 120 hertz causing the extra voltage. As the current through the chokes increases their inductance changes and this goes away, and causes the big voltage swing. If the supply has adequate filter capacity (30 ufd or so) and the voltage swing reasonable without the chokes leave em out. As I recall a choke input supply does have some advantages, but are a lot more complicated than they appear. John
Logged
N4UE
Member

Posts: 709




Ignore
« Reply #5 on: November 28, 2003, 06:55:51 PM »

Thanks John.

Here's some more data. I spent all day today (Gosh, I LOVE retirement!) playing with the beast. Since WB2WIK (and others) have indicated I need very little cap. for the 'tune' cap, (mine is a V V 8-50 pf) I spent some time making a very nice 4 pf (air dilectric) job. I added this in series with the V V. No good.
I had previously stated that the original builder had a 350 pf 'sandwich' in series with the V V. This is incorrect. The 'sandwich' was the DC blocking cap assy.

OK, so the < 4 pf 'tune' cap was a bust. I still get a nice peak with my new 'load' cap. I went back and tried every possible combination I could and the output was still less than when I started. Don't you just hate that? I tried different inductors, etc.

I had put a 1000 pf (5 KV doorknob xmit) cap in there as the DC blocking cap, since that was what was used in every construction article I could find on 50 mHz amps. I got about 600 Watts output. DRAT#$%^&*
I put the 350 pf 'sandwich' back in. Got 800 Watts out. Hmmmmmm. OK, tried a 100 pf cap and got 1000 Watts. Put in a 50 pf(!) and got 1400 Watts. Next I tried two 50 pf caps in series. (these are brand new and the value is verified on my cap bridge) This should give me about 25 pf. It did, and the output dropped to 800 Watts.

At no time during any of this did the output peak before I ran out of cap. In other words, the output was still climbing when the V V cap spacing was maximum.

One other fact. The P/S in question, is indeed a L input. I tried the following experiments:

First, I did a key-down with the inductors (2 in series) in place. The plate voltage went from 3500 to 2000 volts (57%) (at this point, I had 1400 Watts out, 1 Amp of plate voltage and 250 mil of grid current)

Next, I shorted out one choke. The plate voltage went to 4200 no load and 2100 (50%) although the no load was higher, the HV under load was almost proportional.

Next, I shorted the other choke. 3500 and 2000 VDC (57%)

Finally, shorted BOTH chokes. 4800 (!!) dropping to
2600 (54%)

Since the P/S has variacs to control the PV, I could crank in any value, to the detriment of the tubes....

I painted the inside of the RF deck, flat black around the tubes and ensured the blower had fresh air intake and the chimmneys, etc, were in good shape.

The watt meter is a KLM that had it's calibration against a Bird digital ($$$) watt meter from my former job site.

My questions are:

Why am I loosing so much HV? I need to put my cap bridge on my P/S caps to see how much capacity I actually have.

More important: how can I get the V V to peak at mid travel? I do NOT want to buy another. The RF deck is almost built around this baby.

Should my grid current be close to 800 mils? How can I get more? The radio is a 756PRO 100 Watts. The input SWR to the tube is 1.1:1 (been adjusted, LOTS)

Why does changing the value of the DC blocking cap have such a dramatic influence on output? Should I put a variable here. get max output and substitute a fixed value?

Lots of this defies logic. Welcome to my world......

Any suggestions appreciated.

BTW, I operate the amp throttled back to 1000 Watts out.

ron

N4UE
Logged
N4UE
Member

Posts: 709




Ignore
« Reply #6 on: November 30, 2003, 06:12:50 PM »

More of my "answers to my own question".......

Got Big Bertha running the way I expect it to. It took me 8 hours of non-stop work today.

High Voltage problem. I replaced the HUGE mass of hundreds of diodes in the rectifier section with 4 of the "Silicon Alley" 100KV pieces. Mounted them to a very nice heatsink picked up at Dayton for cheap. Even have a fan blowing over it if needed. Voltage drop now 3200 V to 2500 V. Cool......Still using the Choke input configuration. Actually, the choke is in the B- leg, as recommended in older Handbooks.

Plate cap problem. I didn't have enough minimum capacity (sounds wierd!). Tried all kinds of 4 pf homemade Teflon caps in series with the Vacuum Variable, nothing was working. Discovered all I needed to do was massage the tank coil a bit with a 2' teflon rod. The coil is silver plated strap about 3/8" wide. Found a position where the Vacuum Variable peaks at 1/2 travel.

This old girl rocks! Running very easy, I get 1000 Watts output, into the Cantenna. Boy, does that mineral oil get hot quick! Tubes hardly red.

I improved the air system, and painted the inside of the RF deck flat black near the tubes.
On the air reports are just excellent.

One last thing. When I was finished all the above modifications, I was getting an impulse noise that was present when the HV was applied. At 3000 V, it was about once per second, the speed increasing as the HV was increased. Looking at the noise on the PRO, it looked just like ignition noise. I searched for that noise with a 'sniffer coil' for an hour. On a hunch, remembering some of the problems I have had with Power Line noise, I pulled the cover off the end of the 'pole pig'. One of the UNUSED HV taps, had broken free from the mounting screw/insulator assembly. The wire was very close to the screw and was arcing in prorortion to the HV I cranked in......

Now it's ready to work. Anybody need EM80 on 6m?


ron

N4UE
Logged
WW1Z
Member

Posts: 35




Ignore
« Reply #7 on: December 01, 2003, 02:36:44 PM »

The comment about the voltage going up as you short out the chokes sounds like what should happen. Replacing a mess of rectifiers and associated parts with some quality HV rectifiers is always a good idea. At this point you might check your primary voltage loaded and unloaded and see if there is a bad connection there. Certainly check the P/S capacity. Built with those big chokes it may not have enough without. 25 to 35 ufd is about right. Sounds like your getting some where with your amp. Been trying to get some 4-400s going for 6. Not sure which happened first the last time I was working on it, did the dummy load fail and then the big flash and bang from the amp? Or was it the big flash and bang that fried the dummy load? I lean towards the former but with 4KV it could have been the later. John
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!