The AM Forum

Sorry another newbie question... What is the correct solid state replacement for the 5R4GY? Reason I ask is because I picked up a few solid state rectifiers that tubedepot.com sells, to solid state my DX-100. I am able to replace the 5V4 with one without an issue, but if I replace the 5R4GY's with them, it blows 5a, up to a 10a (i didnt try larger), fuses as soon as I key the rig. They say they are the replacement for both the 5R4 and 5V4, as well as a host of others. I also tried using just one instead of 2 in case I was doing something wrong. Also, the rig had a bad LV transformer that I just replaced and it had the issue before and after replacing it. Not that it should matter but just mentioning that. So am I using the wrong one or is there an issue somewhere else? Also, it basically welded my PTT relay contacts together at one point so I also tried just flipping the plate switch, to see if the PTT circuit was causing this issue, and it still happens... Everywhere I look online for info just says " i solid stated the supplies", "make sure to solid state the supplies", or "replace the rectifiers with solid state replacements".

Again, any help for my newbie behind would be greatly appreciated...

Perry
W3MMR

Sometimes one must ask themselves   " WWTD ? "

https://www.amwindow.org/tech/htm/dx100.htm

KLC

I homebrew my own SS replacements:


Phil - AC0OB

Perry,

Check the wiring of the 5R4 sockets against the schematic. If pin 2 from one is jumped to pin 8 of the other, the 5R4's would still work but, the SS replacements could short (and blow the fuse). The wiring of the two sockets should be exactly the same, pin for pin.


Don

1N1239 & 1N2632

The 1N2632 is the best choice.

The solid state rectifier could be bad, even if it is a new one.  I would check it with an ohm meter.  I bought supposedly
new ones at a hamfest and they turned out to be shorted.

It doesnt say really much except "SS the supplies" and what the rating per leg should be. That doesnt tell me much.

Don-- I will check the wiring

Jim-- Thank you.

Oh, and it would still short if only using 1 SS rectifier?

And these SS jobs i have, have 2- 1n5408G diodes inside connected as so: anode of 1 diode goes to pin 4, that diodes cathode is soldered to the other diodes cathode, which is connected to pin 8, and that diodes anode goes to pin 6.

Don-- The wiring of the sockets are jumped to their respective pins. Pin 4 to pin 4, pin 6 to pin 6, and so on.

I check those things with two or three 9 volt batteries in series using a voltmeter.

Perry,

As a test, you could try each SS unit in the 5V4 slot to make sure they're all good. Also, it's a good idea to install a separate fuse in the primary of the low voltage transformer to protect it from overload. I use a 2 amp slow blow fuse in my DX-100. It's cheaper/easier than getting a new LV transformer.

It may be possible to use only one SS unit in the 5R4 slot if it has a sufficient current rating. You'd have to see what you have.


Don

I have several suggestions....

per KLC regarding WWTD, the Tron has done a detailed write-up as linked by KLC....  Toward the end of the document, there is a detailed guide for solid-stating the power supply of rigs such as the DX-100 and Viking series.  First, as he states, the surge when the field of the filter choke collapses on turn-off needs to be tamed with the resistor and capacitor in series, placed directly across the inductor, such that this collapsing surge does not damage the silicon rectifier stack.  The original vacuum tube rectifiers are immune to this surge, but the solid-state rectifiers can be damaged quickly.  Next, consider the reverse voltage requirement (total PIV) of the group of rectifiers in series.  HLR indicates that this should be 7 to 10 KV for the 700 to 800 volt supply using choke input.  I do not believe two of your diodes in the replacement module meet this requirement.  When using diodes in series, they will not each drop the same voltage, so err on the side of MORE DIODES, rather than less.  This may not be related to the cause of your overload at turn-on, but it is necessary to conservatively rate the diodes for long reliable operation.

Your turn-on surge failure may just be related to trying to charge the capacitor quickly with the low resistance of the diode string, as compared with the forgiving high resistance voltage drop over the tube rectifiers.  You may need series resistance between the diodes and the load to tame this surge.  The filter capacitor appears as a dead short before it starts charging.  Read all of this post before proceeding, then if you assume the turn-on surge is the likely failure cause, try bringing up the HV supply on a variac, and monitor the DC to see if it will come up "gently".  Be sure to consider the number of diodes (referenced below) before trying this test.


Since you say the two sockets are wired identically, and the diode anodes are pins 4 and 6, and the cathodes are pin 8, this puts both tube replacement modules directly in parallel.  While the resistance of the tube rectifiers allows them to share the load rather equally, the low resistance of the silicon replacements tends to cause one module to hog the load, because one has slightly less resistance than the other.  So putting both modules in, without equalizing resistors on EACH tube socket, has negligible benefit, as only one module is carrying the load.  It is likely that one is sufficient for the current requirement, but NOT the reverse voltage in this application.  More diodes in the string, and equalizing resistors!

Regarding testing diode MODULES with three or more 9 volt batteries in series, I certainly agree.  But you may wonder why?   A typical DVM or VOM has a very small voltage supplied to the device under test, when testing diodes, transistors, etc.  A higher voltage may damage a signal diode or transistor, so the voltage is intentionally limited.  But  if you do not have 0.5 to 0.7 volts PER JUNCTION, when testing a string of diodes in series, the meter may erroneously indicate an open circuit, when the devices are actually functional.  This is almost always true of microwave oven rectifier modules, which have a large number of diodes in series to withstand the high PIV.  You must overcome the voltage drop of all the diodes in series before you will see considerable forward current when testing diodes in series.  However, if all the diodes are exposed and may be tested individually, then a typical DVM should give you an accurate indication of the condition of each device.

GL, and hope this is helpful to you.....

Another thing before I forget, I spoke to Tim via email last week, and Tim never answered this question, is after switching from cap input to choke input filter, this is before solid stating the HV supply mind you, i have to turn the drive all the way up to barely obtain 5mA of grid current the manual recommends. Is this why Tim says I must SS the HV supply? Or is that the max that im going to obtain no matter what beings that the LV is down around +250vdc as opposed to upwards of 420vdc when cap input filtered? This just seems like a lot of trouble for what gain? Less heat? My 100 runs cool as it is. I can see the benefit of the choke input filter lowering the low voltage so I can use a relay and obtain the screen voltage for the modulators directly from the LV supply as Tim recommends to lower distortion more. Ive already done most of his audio mods and im very happy with the way my rig sounds right now just with a D104 plugged directly into the mic jack, no processing. Over the air using a WebSDR to record, im very happy with it. Using my Anan to record, i can hear the distortion I cannot hear using the WebSDR beings im almost 60 over s9 into the Anan 200d. So I guess im wondering if "the risk is worth the reward" for lack of a better term. And thanks for the patience guys and I apologize for all the newbie and seemingly stupid questions. The boat anchor stuff is still new to me, before this venture, i hadnt used tube theory since A school in the Navy so im knocking the rust off. But having to basically fully restore my DX-100 before using it, it was in rough shape when I got it, has given me a crash course in things. Again, thanks guys.

And heres a recording taken last night if anyone cares to listen.. Lots of QRM, but not shabby beings i started with a non functioning, stock DX-100 2 months ago.

https://youtu.be/RXYBqnBRUvQ

https://www.mouser.com/ProductDetail/Rectron/HVM12?qs=sGAEpiMZZMtbRapU8LlZD0IZEQvpG%2F8RJnSZSuW1xB0%3D

Perry,

I made the same Timtron modification to the LV power supply and had the same problem - low grid drive (especially on the higher bands). I replaced the 12BY7 and the 5763 to no avail. I re-peaked the circuit with the slug-tuned coils and still had low grid drive.

This is what I did to solve the problem. There's a 6.8k 7W resistor in series with the 25k drive pot that goes to the LV rail. I changed the value of the 6.8k 7W resistor to 4k 10W and added a 2k 10W resistor from the other side of the drive pot to ground. This effectively moves the voltage range that the drive pot provides to the 5763 screen grid UP a few volts. That's all it took, worked like a charm.

I did this just a few days ago and was going to mention it to Tim, but he's out of town.

I think you will need to solid state that LV power supply, 250V sounds a bit low. I get about 300V on mine, so do that before changing those resistors.


Don

I did SS the supply, but this was before I blew up the old LV transformer last week. I  haven't tried with the new Xfmr in there. The old one, which I didn't know at the time, was weak, it would heat up instantly, and as soon as I swapped transformers, my peaks went from %130 at best, to easily over %150, my peak power went from 550 at best to 700+ when pushed. So ill definitely give that a try. Thanks Don!

Remember to put a 2 amp fuse in the LV primary. It'll save you the headache of replacing that transformer twice!

Hear Hear !!

Perry,

I think your results, and thus, your question, is a matter of putting the cart before the horse. The reason for reconfiguring the filtering on the PS is because when you solid state the supply the B+ will be substantially higher. By reconfiguring the filtering to choke input you reduce the B+ to approximately where it was prior to the solid stating. The solid stating not only reduces the heat, it also deloads the transformer.

I hope this makes sense.

Scott

I believe that replacing the vacuum tube rectifier with solid state diodes will increase the voltage by a small amount. Going from capacitor input filter to choke input filter will reduce the voltage by a large amount. In the end, the DC voltage will be significantly lower than the stock value (360VDC according to the manual). On my DX-100 the no-load voltage is about 300VDC. That's why I modified the drive control circuit. YMMV


Don

 Regarding the LV supply, also realize that the ripple reduction from the choke input filter will not be as much as a Pi filter. Doubling, or quadrupling the output capacitor value will help. Also the diodes used have less than instant on/off switching times. That means the diode(s) turning ON will occur when the other diode(s) have not turned OFF yet. For a few microseconds each cycle, the transformer sees a dead short. When the slow diode finally turns off, a high amplitude back EMF spike occurs. This spike goes straight through the filter choke's turn to turn winding capacitance, and is also not attenuated by the filter capacitor's series lead inductance.  The result is the low level RF and AF stages fed from the LV supply may pick up a nasty 120 hz buzz There are ways to deal with this, such as R-C snubber across the LV transformer secondary, and an additional R-C B+ decoupling where the 'C' is something like a .1uf 630v polypropylene and the 'R' somewhere around 100 ohms.

"Solid State the Power Supply"
That is right up there with a Chilton's car repair manual that says:
"Remove Dashboard"


Jim
Wd5JKO

Jim,

Interesting. I hadn't given it any thought before. Let's see.

A half wave at 60 cycles is 8333uS and a typical standard reverse recovery time is perhaps 2 uS. So while the conducting diode starts to recover when the forward voltage on it drops to about 0.8V, and the non-conducting diode doesn't begin to conduct until its forward voltage achieves 0.8V, an "overlap" time exists near zero volts. Is this not enough time for the conducting diode to fully recover?

Moreover, with the voltage during this time being so low, wouldn't the current generated by any "short circuit" be very small, especially given the higher resistance of the windings of an older transformers designed for vacuum tube rectifiers?

Additionally, I would have thought that any spikes would have been caused by the relatively fast switching times of the diodes, not a conduction overlap.

Calculating this is beyond my limited math ability, and measuring it is beyond my limited test equipment's ability. And removing my dashboard is just out of the question. ;D


Don

The problem arises due to the leakage inductance of the transformer. A momentary short suddenly going open unleashes a nasty high amplitude transient as the leakage inductance does it's Fury. As mentioned earlier in this thread, a 'R-C' Snubber across the filter choke helps contain the back EMF pulse when the choke field collapses. The old Bill Orr W6SAI radio handbooks talk about this stuff, and gives guidance on how to estimate the R and C.

I learned the hard way early in my career. At work, we had a brute force 30VDC 200 Amp supply. It used a transformer around 70VCT, with a 208VAC primary. The rectifier was 2 big stud diodes with braid on the other end, and a big choke for a choke input filter. The transformer had a lot of leakage inductance because the primary had to be isolated from the secondary to handle 150,000 volts (yep, 150KV). Tranny had layered Mylar and was immersed in oil. We kept losing those diodes just by turning on the power supply. Those big diodes had a reverse recovery time of almost 1 millisecond! That means 1ms out of 8.33ms (60hz) the transformer was shorted out. The cure was a simple series R-C snubber across the transformer. As I recall, it was like 39 ohms 1 watt, and 2.2 uf 100v Mylar Cap. The problem was gone!

More recently I solid stated a Central Electronics 20A Phasing SSB transmitter. Went choke input too. Voltage was a little lower, but I had to fight the audible buzz on the signal, and also noticd the carrier suppression was degraded from the power supply ripple. This can and does happen.

Jim

Jim,

I haven't run into an extreme case, as you have. Typically, I add filter C because I have many salvaged caps in the junk box. Maybe I should start soldering in some of my salvaged MOV's and adding snubber networks too. Sounds like cheap insurance.


Thanks,
Don

Well thats sure a lot to take in lol. Before solid stating, LV was 380-390vdc. Book says it should be around 360. So with my higher line voltage, its right in line. After solid stating, it bumped the LV up to 430-440vdc. I noticed that when moving from cap input to choke input, it brought the voltage down much lower than before solid stating, around 280vdc. So I dont know if bringing the voltage back to where it was originally, was behind the switch. Tim says "This should provide lower screen voltage to modulators and lower voltage to VFO tube. Ideal screen voltage for the 1625 modulators is between 200 and 250 V. Around 250 volts for screens will work very well." So i reckon thats the reasoning for the switch. Anyway, this rig has been running so darn good since I replaces the LV supply, i think im just going to leave it. With exception of a 3 diode limiter just for safety, not for hard limiting. After doing a lot of timmy's mods, im seeing 140% positive before even hitting 100% on the negative, low distortion, the thing runs cool (with a temp probe floating in the center of the case, it doesnt go above 110 F even after a 4 minute transmission), and guys are having a hard time hearing the difference in this and in my Anan 200D and im just using a D10-4 plugged into the front with no processing, as opposed to an ANAN with EQ, multi-band compression, and an SM7B. But im the type to not sit on my hands and "Fix it till its broke". I dunno, you guys tell me... Its not great, but not bad for a 60 year old transmitter and a 40 year old microphone...

https://youtu.be/RXYBqnBRUvQ (https://youtu.be/RXYBqnBRUvQ)


Thanks for all the insight gentleman...

Perry,

Last time I checked, the stock DX-100 modulator screen voltage came from the center-tap of the HV bleeder resistor. That puts about 375 volts on those screens no matter what your LV supply is.  If you want to take screen voltage from the LV supply, you'll need to switch it with a relay or something. Was that part of the mods you made?


Don


P.S. Nice video. That's the best I've ever heard George. ;D

    Haha, he snuck in there! Don't get me started on George. >:( >:( Ive probably told him 6 times in that last couple months to LEAVE IT ALONE AND DON'T TOUCH IT, after helping him to get it to sound decent. And he'll come on the next day sounding like butt, folding his carrier back, so on and so forth. But, i know im not the first to fall into his trap! Still love the guy though. I think its great that he's running an Isotron and hits me at 20 over at times.

    Anyway, back to the topic at hand. If you read his "DX-100 Mods" page, theres many parts in there where he jumps around and things get confusing. Before the text I had quoted, he says this:
 
"RF Final Screen Supply

The screen voltage is derived from a divider in high voltage power supply. This is a poor method. Under normal operating conditions, screen current will cause that screen voltage to sag. In the original circuit, there is a large, 30KW, center tapped resistor, which establishes a half voltage point for the two electrolytics in the power supply, but under normal operation there is an imbalance of voltage, and the electrolytic from the top end of the hi-voltage supply to the center tapped resistor will wind up with over 450 volts across it, boom! - the electrolytic will break down! The screen voltage actually needs to be lowered. The screen voltage should be taken directly from the low voltage supply, through a relay. This is done when the PTT circuit is added. For example, a 4-pole relay is used to key up plate voltage, exciter B+, and modulator screens, as well as audio driver B+."

    The way he puts it, he makes it sound like hes talking about the screen voltage for the 6146's, to my novice behind anyway. I guess because its labled "RF Final Screen Supply" The way he put things in his post really confuses me. And took a long time to weed through. So ive gone with other folks' interpretation of the mods.

    Moving on. Looking through my notes, I have 349vdc on the center tap of the 30k bleeder resistor. What i did was add another 30k of resistance and it brought the screen voltage down to 278vdc according to my notes. The final grid bias I settled at is -32vdc to get the mod transformer idling around 70mA as Tim recommends. Now, I know that adding resistance isn't the correct way to lower the screen voltage. I had tried adding a relay and picking off the LV supply at one time, after switching to choke inpoot. My results weren't good. I dunno if the relay was not stout enough, but the audio was breaking up and very distorted when I tested it, so I added resistance instead by using the bleeder resistor off of my parts rig. I havent noticed any ill effects on anything from doing it that way, but maybe you'll have some insight on that. One thing I havent done is add negative feedback. I wonder how much of an improvement adding a negative feedback loop around both the 12AX7 and 12BY7 would be as far as reducing distortion. Ive read where guys have done this and it didnt really make any difference, so they took it out.

Thanks Don, and Happy Fathers Day.

Perry,

Interesting. The common wisdom is to use a regulated voltage on the modulator screens for the best results. A series dropping resistor would give rather poor regulation. But if it works for you, then it works.

In stock form, the mod screens have a .1uf bypass cap hung on them. Did you add any more C there to try and hold the voltage up? Maybe a few hundred uF? That might be something to try.

I went through those kind of mods some months ago and ended up replacing the 1625's with 6DQ5's and driving audio to the screens from an external amp. No screen supply needed, just a zener for bias. I'm working on building an internal audio driver but that's moving slowly.

I haven't played with NFB so I can't comment on that.

For me, it's the process that's fun. Trying things, learning things, and making the Heathkit DX-100 into My DX-100. You can do the Timtron mods, put all the screws back in the cabinet, throw your soldering iron away, and be done - you'll have a very nice rig that way. But, if you sharpen your machete, you can try to cut a new path through the jungle and call it your own. Admittedly, many have done just that. There's hardly a tree left to fell in the DX-100 jungle. The journey is the best part. You've joined a long tradition of hams modifying their AM transmitters. I hope that you enjoy it as much as I have.


Don

Perry,

It is likely that George goes on SSB and cranks the audio, uses the processor.
Then forgets where he had it set for AM...
I think it is that simple...

Bear,
I didnt think about that. Good point. He's been getting a lot of compliments lately and guys saying its the best theyve ever heard him. He was up on 40m the other day sounding great. Happy fathers day to you as well bud.


Perry

I have 5R4 tube bases and SS diodes, and could put together a couple 5R4 replacements. I understand how to connect the SS diodes to the tube pins, but do you use shrink tubing, any kind of enclosure? Any photos?

Brad

I just daisy chained the diodes on an octal base and covered the devise with a pill bottle. Fits better than OJ's glove.

klc

This is why I use pc based outboard processing.

With a mouse click, I have ssb wide, ssb for DX and am settings.

Makes life simple, although getting the rf out of the virtual air chain wasn't - - Shane
KD6VXI

Brad,
What diodes do you plan on using? These are the ones I was looking at. https://www.mouser.com/ProductDetail/Rectron/HVM12?qs=sGAEpiMZZMtbRapU8LlZD0IZEQvpG%2F8RJnSZSuW1xB0%3D


Thats exactly why I DONT use my external rack with my Anan anymore. It automatically switches transmit profiles when I change modes, bands, whatever I have it set to do.



Sounded good down this way last night Don. Wish it wasnt as crowded so we could have had a QSO, but hey, thats the AM window @ night in the Northeast.


Perry

Yes, it is. And I wouldn't change it for the world.  :)

Perry,

I don't know why the diode you linked to wouldn't work. I forget what the current flow is at the HV rectifiers in the DX-100 PS so make sure the diodes you buy are appropriately rated.  I'm thinking I will use 1N5408 as I have them on hand.  I think they are 1KV jobs rated for at least two amps but I will have to check that.

I'm interested in SS rectifiers more as a trouble-shooting tool.  When I was working on my DX-100B, SS rectifiers in the HV supply would have allowed me to find a couple of problems more quickly at a much lower voltage and probably without the smoke.  :)

Brad

Sorry, no pictures but I did use plenty of shrink tubing to and from the "tube saver" octal base. Since I had a cabinet surrounding the unit, I did not enclose the "tube saver" octal base.

I essentially did what WIG did and simply connected the 4, 1N5408s at the base and then to the resistor and then ran that to the input of the choke.

R* may have to be adjusted for your specific transformer voltage and how much HV B+ you want to allow.


Phil - AC0OB

hmm weird. Anyway, well Brad, the 1N5408's are whats in the SS rectifiers i already have, that are wreeking havoc when I put them into the HV supply. As soon as I key the radio, it blows the inline fuse I installed. I use a 6a fuse in normal operation, so i tried a 10a, still blew, so I knew something was up. At one point, it even welded my PTT relay together in the TX position. So just a heads up. On the other hand, if they end up working for you, then I know I have something else going on. I also tried just one of the rectifiers at a time, as opposed to 2, in both sockets to see if that would work and I still got the same result. So be careful, and let know how how ya make out! I'm very very curious to see.


Perry
W3MMR

Phil,
 The SS jobs i have, have 2- 1n5408G diodes inside connected as so: anode of 1 diode goes to pin 4, that diodes cathode is soldered to the other diodes cathode, which is connected to pin 8, and that diodes anode goes to pin 6. You said you used 4 diodes in each and it works fine? Are you using both sockets or just one? Thanks!!!


Perry
W3MMR

Perry, I used 4 diodes total (as per the schematic, Sheet 1) for a bit of extra reverse voltage headroom.

It is only necessary to use one socket as the HV supply current is around 350 mA.

Use at least #16 stranded, insulated, and shrink tubing cover for the wiring to and from the socket and resistor.  

Phil - AC0OB