Tube Rectifier info and solid state replacements

[WG9P]

I have a Collins 32V2 that has gone through two low voltage transformers. Consequently I have decided to install solid state rectifers in order to lower the heat and current draw on the LV xfmr.
In my research I have found lots of plug in solid state replacements. However, I know I can build a better product for pennies. I am worried that all these so called plug in replacements do not provide a correct series resistance to simulate the voltage drop of the rectifier tubes.
This link: http://www.fourwater.com/files/fullrect.txt is full of information on just about every tube rectifier I have come across and also describes how to go about designing and building a solid state plug in rectifier for your particular application. In my case I am using the 5Z4 for the low voltage (240VDC) and 5R4 for the 700VDC HV circuit. I will use two 1n4007 diodes and a 50 ohm resistor in the 5Z4 device and two 1n4007 diodes and a 150 ohm resistor for 5R4 replacement device.
I built one of these for my 75A-3 to replace the 5Y3 using a 50 ohm resistor  and it works just fine. As a matter of fact I immediately noticed the hum that has plagued this receiver almost entirely gone now. schematic and photo attached. I know that there are purist out there that think it is blasphemy to do this but I think it is a wise choice to reduce the strain on our favorite old rigs so they can continue to run for years to come.

[ke7trp]

Thanks for the link.  I think the SS rectifiers can be a good choice as long as you use that resistor, measure the ripple and adjust the capactiance so you do not add hum to the rig. 

On my transmitter, Running SS replacements for the 5R4s and 3b28s removed a staggering 16 amps of current and a very large amount of heat. I can feel the difference in the small ham shack just in heat load.  Also the temp of the case of the transmitter is much higher.  To bad, All those supplys had increased ripple, that turned up on the output of the rig.  I switched them all back to factory.

The last thing you might want to measure is the inrush to the transformer when you turn it on with the SS.  I am thinking that the resistor is going to help there.

C

[KA2DZT]

I've built many high voltage supplies and never used any series resistors.  Not sure why you would use them.  I guess to mimic the original voltage drop of the tube rectifiers.

I would suggest that you use at least two diodes in each leg of the higher voltage supply.  I use 3amp 1KV PIV diodes (1N5408).  I usually use four diodes in each leg of any supply from 600-1000 volts.  I like a big safety margin on the PIV.  All my supplies are choke input filters, this limits the in rush current.

There has been some reports of higher hum levels with SS diodes.  I'm not sure why this would be the case.  I would think the higher hum levels (if they are higher) are caused by some other factor.  Maybe the cause is poorly designed filters.  Input chokes that are not up to the task.  I always use two stage filters (LCLC) and my hum levels are down to near nothing.

To repeat,  I strongly suggest that you use at least two and better three diodes in each leg of the 700 volt supply.

Fred

[RolandSWL]

Interesting reading........

http://www.w8ji.com/power_transformer_stress.htm

Roland..............

[KM1H]

You beat me to it Roland and Ive cited it on here several times whenever the SS discussion comes up.

If you simply replace a 32V transformer with another you will keep burning them out as they are poorly built with cheap insulation. OTOH having one rewound is the way to go and leave the tube in place.

I certainly would not use 1N4007's in the HV circuit in any case, the 1N5408 is a much more robust choice.

If you truly want to reduce heat and increase reliability then use a bucking transformer on the primary. Getting down to 108-113V will make a big difference and maybe even save the rest of the cheap iron. And, yes I have a 32V2 also and follow my own advice.

Carl

[ke7trp]

Great read roland.  Thanks for posting. 

C

[KZ5A]

You have to take W8JI's articles with about a tablespon of salt.  He does present some very good information at times but his main focus seems to be explaining why anything someone else believes is wrong.  I got so tired of his pontificating on the Elecraft reflector that I dropped off of it.

Bottom line is that if you remove some of the load from a transformer it runs cooler.  

A big benefit of replacing tube rectifiers with SS diodes is that the 5V filament windings can (and should) be disconnected from the tube sockets.  This eliminates one of the most common failure modes for vintage transformers.

In circuits where the extra voltage might be a problem, the series resistor restores the original voltages.   However in most cases, particularly in the HV supplies, the slight extra idle voltage does no harm and may be beneficial.   The dramatic improvement in voltage regulation, driven by the elimination of the "effective series resistance" of the tube rectifier, is a big plus, particularly for modulators and linears.  

As we speak, I'm in the middle of solid-stating the rectifiers in my B&W 5100.   In addition to the rectifier change I'm adding a Harbach soft-start module to the HV primary.   I'm also adding a thermistor to the main AC to provide some soft-start for the entire TX on power up.  I'm also increasing all of the filter caps to around ten times the original "C" values.  

The LV/Bias supply  in the Barking Willy is a squirrelly design (changed in the 5100B) that may need the "ESR" of the tube, or to have the dropping resistors for the VR tubes re-adjusted, TBD.

The HV rectifier tube sockets are being re-purposed to hold a pair of VR-105s to regulate the modulator screens.

The net effect of all this is expected to be reduced stress on both power transformers, a 90% reduction in HV hum, and a significant improvement to the already good audio.

The only downside is loss of "originality".   My 5100 lost it's "originality" cherry a long time ago and appears to have been modified and re-modified several times in the past, so no worries there.

73 Jack KZ5A

[w1vtp]

Carl

You've made two good points in my view:

1) The bucking transformer idea helps a lot - I believe HLR has recommended this approach with 32V's

2) Rewinding a toasted trannie gets you not only one that will outperform the original but will not change the appearance of the 32V.

Al

[WA1GFZ]

you don't need no stinking bucking transformer. Just put the spare 5 volt winding in series with the primary phased properly to increase the number of primary turns so it isn't sitting close to saturation at 120 VAC. Transformer operating temperature drops and there is less load on it with useless rectifier tubes yanked out. Rig operating temperature drops as well.

[K5UJ]

If the vacuum rectifier does the job and the temp rise on the transformer isn't much, I'd run the vacuum rectifier.  One problem (in my opinion) is that they always get hot as hell and if the parts layout has the tube sitting 1/2 an inch away from the iron it'll cook the side facing the tube.  that's how it is in the 75A-3.   For that if no other reason, I'd go s.s.   

I've been working with a Knight T-50 lately with a 5U4 but it is by itself and good 6 inches from the p.s. iron which feels barely warm even when the rig is key down into a DL.  On that one I stick with the 5U4.   On a CW rig I wonder if s.s. diodes would spike up the leading edge of the CW envelope.

[k4kyv]

Replacing a vacuum rectifier such as the 5U4 with s.s. diodes should reduce the spike on the leading edge CW waveform, since there would be less +HV sag under load, as the filter caps discharge down from the unloaded to the loaded plate voltage.

[K5UJ]

Okay, this is counter intuitive to me--I guess I assumed the vacuum rectifier resistance might cause the leading edge of the keyed CW envelope to ramp up more slowly.  But since it is working through the choke and one or two filter caps I wasn't sure.

[k4kyv]

The envelope of a CW waveform is usually atrocious with a choke-input filter in the power supply, unless plenty of filter capacitance is used and/or the supply is heavily bled down.


Pre-WWII transmitters often used choke input and 1 or 2 mfd of filtering.  But most hams were blithely oblivious to how bad their waveform really was, because oscilloscopes back then were about as unobtanium as HV filter capacitors of more than a couple of mfd, and as bad as the waveforms look on the scope (due to the poor dynamic regulation of the HV supply), it doesn't make all that much difference in what the note sounds like at the other end of the QSO.

[KL7OF]

you don't need no stinking bucking transformer. Just put the spare 5 volt winding in series with the primary phased properly to increase the number of primary turns so it isn't sitting close to saturation at 120 VAC. Transformer operating temperature drops and there is less load on it with useless rectifier tubes yanked out. Rig operating temperature drops as well.

What he said.....+1