EMRFD Message Archive 13828

Message Date From Subject
13828 2017-04-28 09:35:18 Nick Kennedy Reducing line voltage for old tube gear
I recently bought an HT-37 (congratulations to me!) and am checking it out on the workbench as well as gathering tidbits from the web on other user experiences.

With boatanchors, there's always that bit of concern about how line voltage has risen in that past fifty years or so. Additional unloading of the transformer by "solid-stating" the rectifiers can increase voltage even more, with 5 V filament windings unused. One user on the web said he put the 6 V winding of a filament transformer in series with the line. That seemed interesting. Another said he always brings his up with a Variac. That seems extreme.

My thoughts on dropping some voltage in the primary go way back to a P/S I built in the 70s with a little too much voltage from the transformer. As I recall, I calculated how much resistance it would take to drop 10% of primary voltage and put that in the line. But it failed miserably and I decided that it was because I didn't account for the fact that rectifier conduction is typically in short pulses, so the primary current isn't sinusoidal.

But maybe my memory is faulty. I've been doing some models of this issue in LTspice and the "resistor solution" doesn't seem as bad as I remembered.  But it might still dissipate a lot of power and adversely affect voltage regulation as transmitter output varies with keying or voice peaks.

The "filament transformer" idea would fix the wasted power part but still degrade voltage regulation. 

Another idea that occurs to me is to take the unused 5 V winding and connect it into the primary in a bucking configuration. With a transformer designed for directly heated cathodes, the winding should have a high voltage rating. A single 5 V buck winding won't get me 10%, but it's something.

The filament transformer method is like increasing leakage reactance. The buck winding is like changing the transformer ratio.

I wonder what thoughts or experiences others may have on this. Is trying to compensate for higher line voltage a fool's errand?


Nick, WA5BDU

13829 2017-04-28 09:57:57 Clint Re: Reducing line voltage for old tube gear
Hi Nick,
The same problem exists for the Atlas 210X AC console power supplies.  The DC voltage to the TX PAs is unregulated and was very high even when the primary mains were at 110 VAC.
In many years past, I have rewound transformers to change the output voltage, but the effort was very labor intensive.  It was extremely difficult to remove the steel laminations on the quality transformers.
13830 2017-04-28 11:33:27 Rick Herndon Re: Reducing line voltage for old tube gear
Nick, if you do the filament winding buck scheme, remember that winding current rating must be greater than or equal to the rig current draw from the primary. All primary current flows through the filament winding wire. 

Rick Herndon K5FNI

13831 2017-04-28 11:43:18 anglerhamming@out... Re: Reducing line voltage for old tube gear
Nick and the group,
I have a large (25 pounds) Kenyon T-377 filament transformer that I'm mounting on top of an old cash box. I have to put aluminum brackets around the joined edges of the box's sides and front to stabilize the thing. Can I remember when and where I got the thing? No.

The transformer has 5 separate secondary windings, the largest of which are 5 vac at 6 amps and 6.3 vac at 5 amps. The 5 volt winding in buck configuration with the primary gets me to to 117 vac from 122 at my house.  The advantage of a separate transformer is that is that I can use it with my 3 boatanchor transceivers, a Galaxy V Mk2, an HW-101 and my semi-BA Tempo 2020. Since the transformer will now be an autotransformer, the  117 volt secondary of the transformer is not isolated from the AC line, same as the variac that I have that used to be in a washing machine.  The green safety ground wire feeds (or will, when I get this contraption finished) through the box via a standard three prong outlet. This will be feeding the three individual power supply transformers that are still providing chassis isolation to the rigs. I have modified the Mk V and 101's power supplies to three-prong power cords. This setup will *not* provide isolation to my S-38 receivers; I have back-to-back 120-to-12.6-to-12.6-to-120 @ 80 volt-amp rated transformers on a much smaller box for those.

Ted, KX4OM
13832 2017-04-28 12:06:15 John Marshall Re: Reducing line voltage for old tube gear

If you have a spare filament transformer, why use its filament winding alone? If its primary is intact, you can wire it to buck the HT-37’s primary voltage in the same way you would use the 5 volt winding. Or you could use both together for 11.3 volt reduction.

John, KU4AF
Pittsboro, NC

13833 2017-04-28 12:34:27 Ed - K9EW Re: Reducing line voltage for old tube gear
Hi Nick,

What do you think of this approach? Nothing fancy, but it gives you 115VAC
for the primary of your transformer.

[image: Inline image 1]

ed - k9ew

13834 2017-04-28 16:22:34 Will Kimber Re: Reducing line voltage for old tube gear

HI Folks

In the old days of TV sets directly connected to AC mains in UK and else where they used Thermistors. These were 5/10/20 watt negative co-efficient resistors that  started off in the hundreds or ohms and as they warmed up (HOT - very HOT) went down to tens of ohms.

Put in the series heater line controlled the initial surge current and in HT they  reduced the peak voltage until heaters warmed up and valves/tubes conducted.

I believe such things are still available.  Could be put in primary circuit,  probably the best place, or individual secondaries.



13835 2017-04-28 18:02:29 Bill Carver Re: Reducing line voltage for old tube gear
I'm going the same thing, Nick: an external 2x7.5V at 20A to buck the AC line to  ALL  of the old hp test equipment.
Bill W7AAZ

13836 2017-04-28 23:57:36 Bill Carver Re: Reducing line voltage for old tube gear
I pulled the 5Y3 from an HQ-100. I used its filament winding to buck the
primary and it works fine to reduce the secondary voltages by about 4%.
That reduced 6.3V filament voltage closer to 6.3. IE: IT WORKS.

That said, 4% won't compensate for the increase in high voltage that
silicon diode rectifiers produces. In a receiver, with constant load on
the B+ line, I put a wirewound resistor in series with the silicon diode
and that compensated the HV increase fine. In an exciter, with varying
load, a resistor may not compensate for silicon rectifiers as well.

Bill W7AAZ
13837 2017-04-29 05:19:59 Nick Kennedy Re: Reducing line voltage for old tube gear
Is the drop across a rectifier tube constant, somewhat?  In my HT-37, a previous owner "solid stated" the rectifier tube for the LV section and put a 100 ohm resistor in series with the diodes. The HV section hasn't been done.

I probably misunderstood what I'd read elsewhere about use of a filament winding as a series reactive element. Of course it would need to be powered. My initial thought was that passing 4 A throw a winding rated 5 V at 4 A would drop 5 V. But I'm sure it would be much higher than that.


Nick, WA5BDU

13838 2017-04-29 07:43:21 Bill Carver Re: Reducing line voltage for old tube gear
Let's say the primary of a filament transformer is rated at 117VAC. If you power it from a 120 VAC line it'll actually produce 5.1 volts. If you then use that voltage to "buck" the primary of separate transformer the input to that transformer will be 120 - 5.1 = 114.9 volts.

But the idea is to use the otherwide-unused 5V winding to do the same job. Bucking the primary with it will also reduce the 5V winding voltage so you don't get quite as much effect. In this case, if the 120 VAC is applied to the series connection of primary and "5V" winding. A little bit of algebra produces the relationship that the primary winding voltage will be 120 times 117.5 divided by 117.5 + 5 = 115.1 volts AC. So you get 0.2 volts less reduction by using the "5V" winding because it's ALSO subject to the reduction of line voltage to the primary. That 0.2V less effect is not a big deal, especially considering that that 5V winding is "free".

As someone pointed out, the current rating of that 5V winding should be at least equal to the primary current. In the case of my mod to the HQ-100 with a 5Y3, the filament winding is rated at 2A, which would be about 230 volt-amperes (watts, if the power factor was 1). That has plenty of margin.

In the case of the HT-37, I assume it was using something like a 5U4, with a 3A filament. So the "buck" situation should work up to 3A of line current without any problem. It's a thermal issue....heating of the winding. If you are very long winded, and it was an AM transmitter, the 3A with something like a Viking Valiant would not be good. But for a CW/SSB rig like the HT-37 the line current is way below 3A when not transmitting. Peak current might possibly be a bit above 3A but you're not key down or talking more than 50% of the time. so I expect it will be OK. A line-measuring gadget like a "KILL-A-WATT" would confirm that.

The 100 ohm resistor with the silicon rectifiers is a "must" on older gear. It reduces the peak current, just like the old vacuum tube rectifier did. With that resistor the filter capacitors will have peak current and voltage about what it was with the tube. Which is a good since the idea is to reduce total power consumption and heating and reduce the stress level on everything.

Bill W7AAZ

13841 2017-04-29 18:10:10 Bill Carver Re: Reducing line voltage for old tube gear
P.S. Didn't answer your question, Nick.

The voltage drop across a tube rectifier is fairly constant ONLY if your
rectifier is mercury vapor 816/866 (about 15V as I recall) or gas-filled
tubes the military used (3B25/3B28 as I recall). It's a nonlinear thing,
but the usual 5Y3 and 5U4 have effective resistances on the order of 100
ohms. A 1KV silicon rectifier has 0.7V drop, even if you series them for
high voltage their drop is a few volts at essentially ANY current.

A receiver or lower voltage supplies in a transmitter usually have a
capacitor input filter. In those supplies it's necessary to put a
resistor in series with the silicon diodes to keep the B+ from soaring
toward the peak voltage. As you observed, there was a wirewound resistor
in your HT37 for that purpose. It's a nonlinear circuit so difficult to
compute a resistor value ahead of time to produce the same DC voltage as
the original tube rectifier. Hopefully you have resistors in your
junkbox, and can start with a few hundred ohms, reducing it until you
get the original DC voltage.

Bill W7AAZ
13842 2017-04-30 09:10:57 kb1gmx Re: Reducing line voltage for old tube gear
>>>>A receiver or lower voltage supplies in a transmitter usually have a 
capacitor input filter. In those supplies it's necessary to put a 
resistor in series with the silicon diodes to keep the B+ from soaring 
toward the peak voltage. As you observed, there was a wirewound resistor 
in your HT37 for that purpose. It's a nonlinear circuit so difficult to 
compute a resistor value ahead of time to produce the same DC voltage as 
the original tube rectifier. Hopefully you have resistors in your 
junkbox, and can start with a few hundred ohms, reducing it until you 
get the original DC voltage.<<<

Having a HT37 the problem is the transformer is would for 110-115VAC
and the standard now is about 10% higher.  Solution is make it 110V
or risk a very hard to find transformer.   Solid stating the rectifiers only gets 
rid of trivial heat for the 5V rectifier heater windings  and a few ohms is enough
to protect the rectifiers.  In my case the solution was to use the two 5V windings
in series with the 110 volt winding.  It just required lifting the wires and getting 
them phased correctly to lower the input by about 10V.  The alternate is a 
12,6V 3A transformer in a buck configuration.  By lowering the cabinet temp
I found the VFO was less drifty and settled faster (lower temp).

While the resistor in series with the rectifier is for voltage drop in some cases it
is not guaranteed to hold the desired voltage.

The nominal use for the resistor with semiconductor rectifiers is to limit 
PEAK CURRENT and repetitive peak current under loads.  For a 1n4007
that is about 1a continuous and 30A single cycle peak.  The later condition
is meat by assuming the caps are fully discharged (0V and applying the 
full transformer to the diode usually there is enough resistance in the 
secondary to limit to well below the 30A single peak.  For the later case 
the Seconday resistance plus some resistor is there to keep the repetitive 
peak current below 1.4A and average to 1A or less.  Usually that value is
about 1 ohm per Peak AC Volt (at 550V about 550 ohms minus the 
resistance of the winding) at least for the 1n4007 case for the more
robust 3 and 6 Ampere diodes it can be lower.

None of that helps keep the peak voltage down as it can drift up even
with large values if the load is light.

Resistors ruther down after the first filter is for accommodating low power
circuits that required much less voltage or gas regulator tubes.

The best way to limit peak voltage is to limit applied AC.  The usual
solution for old gear is a 12.6V 1 to 3A transfomer on the primary 
side wired to buck the AC to the equipment transformer to 110V 
rather than 120 (or sometimes higher) input voltage.  

Why do that?  Its not about the DC secondary voltage its as much as
preserving the transformer from voltage breakdown.  The transformer 
is hard to replace the caps are easy.  THe following is an example 
for a typical transmitter transformer for the 2x 6146 class.

If a 110/550V (770V peak) transformer is subjected to 120V its secondary 
sees 600V (848V peak).  IF the line voltage were to drift to 125Vac we 
would see 625Vac (or 883V peak) subjecting the insulation to potential 
breakdown voltages.

Add to this the typical ICAS rating for 6146 is about 700V at 110V line we 
are reasonable.  For the 125V case we exceed that easily by 100-150V
and increase the dissipation beyond the max for the tubes or trade life.
Since the tube are working harder the transformer is as well and runs 
hotter.  Both shorten the expected life of named components as well 
as others.

A good example of this is the Tempo-One.  The DC to the finals is 
supposed to be around 700v  typical is much higher.  The 300V supply 
is also higher as well (350-360).  Why?  The transformer is wound for 
110VAC  and the modern 120-125V line is pumped up the output.
The radio tolerate it save for the following.

*The filament is higher than rated, result lower life.
 Nominal is 12.6@110V and for 122Vac it is now 13.9.... too high.

*The plate voltage being higher and no derating of the plate current
means the sweep tubes used are even further stressed and fail sooner.
Normal max power input is 700V *250ma or 175W(DC) but at 122V 
ac line the 800V*.25=200W.  Along with the setup bias at 800V is 
too high for the standard bias setting as a result.  Since the manual
assumes 110V for all setting the result is about 10% high at 122V
and marginally high enough for the tubes (driver and finals) to shorten 
the life further as they are run at over their standard ICAS ratings
at 700v.

* C55 a 500V 100pf molded mica from the driver to the final grids.
 Many would say the part is marginal in vottage rating  But at 110V
 is was within (barely).  What happened?  At 110V the driver tube 
was at 300V and the grids during receive at -120V so the cap only 
saw 420V which is under the 500W breakdown rating.  However 
at 122V the driver plate is almost 360V at receive and the grid bias
during receive is -150V putting 510V across that poor cap which 
fails and puts over 300V on the final grids and the tubes fail in under 
30 seconds sometimes taking the power supply with it. (FYI its 
quite spectacular to see!!)

That is why limiting the input voltage is important.

This is not commonly seen with solid state gear as most run from 
regulated supplies.  Its uncommon for tube gear to have regulated 
supplies save for VFOs.  The problem is tube use step up 
transformers and a 10% change in line voltage can cause 
huge changes in plate voltages and currents with attendant
component stresses.

Long answer but there are details to pay attention to.


13843 2017-04-30 12:39:03 Nick Kennedy Re: Reducing line voltage for old tube gear

​I have really gotten a lot of good information from the responses to my questions. Some additional remarks and notes -

The HT-37 has a 5V4G for the LV (335 VDC) voltage and a 5R4GY for HV (750 VDC). Each of those rectifiers requires 2 A filament current, so presumably that's the rating of the 5 V windings.

The LV section has a capacitor input filter and the HV section has choke input.

Use of the filament transformer to buck the line was of most interest to me. I'm a guy who can't resist a nice transformer  with common output voltages at a hamfest when it's cheap. ​(Or at least common 50 years ago maybe.) So I had a nice new in box Thordarson transformer rated 117 V to 6.3 VCT at 10 A, that I'd paid $1.00 for.

I've hooked it up temporarily and with a dual wattage soldering gun and an embossing hot air tool (for SMT soldering) as loads I got these results with line voltage at 122.9 VAC:

No load:    116.0 VAC
140 W:      115.2 VAC
100 W:       115.6 VAC
320 W:       114.1 VAC

So it looks pretty good. The HT-37's manual says input voltage can range from 105 to 125 VAC, but for a 50+ year old rig, why push it on the top end? I'd appreciate the same consideration myself. ;^)

In my rig, the LV supply has been "solid stated" but the HV has not. I'm considering whether to do the HV.  Meanwhile, I noticed that the hookup of the existing modification left the 5 V winding still connected to the 335 VDC output, even though the winding no longer supplies current. No harm maybe, but I decided to completely disconnect it. While doing that, I found that the 100 ohm resistor was connected incorrectly and was not in series with the output at all. The things you find! So I fixed that while I was at it.


Nick, WA5BDU

13844 2017-04-30 15:25:42 Hubert Miller Re: Reducing line voltage for old tube gear
Dealing with receivers, the higher quality ones usually had a filter choke. If you simply change the filtering to choke input from
capacitor input, by moving the input capacitor to in parallel with the second capacitor, you drop some more voltage. Tube 
circuits, at least for receiving, are quite tolerant of lower voltages. In fact, the National Co. manuals i have looked at, specify
for battery operation, B voltage of down to 135 volts, this from around 200-something when AC powered. The choke input
will likely not compensate totally for the higher AC line voltage today, plus the 25 - 35 volts gained by using s.s. diodes, but
it means you can use lower value B+ adjusting resistor, which means less heat generated. 
13845 2017-04-30 15:38:31 AD7ZU Re: Reducing line voltage for old tube gear
Good points Mark,

HT37 HV P/S (+700) uses a choke input filter 
The LV supply (+300) uses a capacitor input pi filter.