16 Jul
2025
16 Jul
'25
7:13 a.m.
Recently I got a nice and complete PDP-8/s from the US. The power supply
uses a ferroresonant transformer which in addition to the standard primary
and secondary windings has a separate 2.3H winding connected in series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The transformer's
secondary side and the resonant circuit are operated in saturation. There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to the
input frequency of 60Hz. This is quite a nice if not elegant design for the
period in question, but maybe not the most efficient.
As I live in Australia I get 240VAC and 50Hz as opposed to the US 115VAC
and 60Hz.
I can easily convert our 240VAC to 115VAC with a step-down transformer, but
cannot easily supply 115VAC at 60Hz. So I was considering using a step-down
transformer to get the 115VAC, but modify the resonant "tank circuit" for
50Hz.
Unfortunately there is some magic I don't understand. The resonant
frequency of a LC circuit with L=2.3H and C=2uF is about 75Hz not the
expected 60Hz.
Otherwise I could just solve the standard LC resonant circuit formula for C
and plug in 50Hz and 2.3H to get the required C.
Obviously ferroresonant transformers are more complex than this former
software engineer can grasp. Could any experienced EE with relevant
transformer knowledge please chime in and help me understand how to
redimension the tank circuit to use 50Hz instead of the original 60Hz input.
Obviously I could replace the entire power supply with two modern switch
mode supplies to create the two rails, but it would be really nice to keep
the original supply and just reversibly adapt it for 50Hz.
Thanks and best regards
Tom
16 Jul
16 Jul
7:43 a.m.
On 7/16/25 09:13, Tom Hunter via cctalk wrote:
Recently I got a nice and complete PDP-8/s from the
US. The power supply
uses a ferroresonant transformer which in addition to the standard primary
and secondary windings has a separate 2.3H winding connected in series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The transformer's
secondary side and the resonant circuit are operated in saturation. There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to the
input frequency of 60Hz. This is quite a nice if not elegant design for the
period in question, but maybe not the most efficient.
As I live in Australia I get 240VAC and 50Hz as opposed to the US 115VAC
and 60Hz.
I can easily convert our 240VAC to 115VAC with a step-down transformer, but
cannot easily supply 115VAC at 60Hz. So I was considering using a step-down
transformer to get the 115VAC, but modify the resonant "tank circuit" for
50Hz.
Unfortunately there is some magic I don't understand. The resonant
frequency of a LC circuit with L=2.3H and C=2uF is about 75Hz not the
expected 60Hz.
Otherwise I could just solve the standard LC resonant circuit formula for C
and plug in 50Hz and 2.3H to get the required C.
Obviously ferroresonant transformers are more complex than this former
software engineer can grasp. Could any experienced EE with relevant
transformer knowledge please chime in and help me understand how to
redimension the tank circuit to use 50Hz instead of the original 60Hz input.
Obviously I could replace the entire power supply with two modern switch
mode supplies to create the two rails, but it would be really nice to keep
the original supply and just reversibly adapt it for 50Hz.
That L is probably different when the transformer is excited
at full mains voltage and near saturation. Of course, this
seems like it will reduce L and therefore drive the
resonance higher!
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up. Then, measure the output
voltage, and if it is within range just use it like that.
If you want to get fancy, put it on a Variac and sweep the
input voltage. You will note a reverse slope, as input
voltage rises through the range, output voltage will decline.
Jon
8:10 a.m.
On Wed, Jul 16, 2025 at 4:00 PM Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up.
Remember that the resonant frequency goes as 1/sqrt(L*C). So doesn't
that mean you want a capacitor of 1.2^2 times the original value?
-tony
8:42 a.m.
On 07/16/2025 11:10 AM EDT Tony Duell via cctalk
<cctalk(a)classiccmp.org> wrote:
On Wed, Jul 16, 2025 at 4:00 PM Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
One thing to keep in mind is that the transformer may not be designed to work at 50 Hz. A
"normal" transformer will saturate if the frequency is too low. A lot of 60Hz
equipment won't work (well) on 50 Hz unless the transformer is specifically designed
for that. However, how that applies to a transformer that is already intended to saturate
I don't know. I do suspect it is still an issue since it will likely saturate sooner
than expected, for some definition of sooner.
Will
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up.
Remember that the resonant frequency goes as 1/sqrt(L*C). So doesn't
that mean you want a capacitor of 1.2^2 times the original value?
-tony 
11:45 a.m.
On Jul 16, 2025, at 11:42 AM, Will Cooke via cctalk
<cctalk(a)classiccmp.org> wrote:
That's true but the frequency difference is modest, so chances are it will be ok
unless the design is really marginal. It would be a different matter if you were dealing
with a transformer designed for 400 Hz power.
paul
On 07/16/2025 11:10 AM EDT Tony Duell via cctalk
<cctalk(a)classiccmp.org> wrote:
On Wed, Jul 16, 2025 at 4:00 PM Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
One thing to keep in mind is that the transformer may not be designed to work at 50 Hz.
A "normal" transformer will saturate if the frequency is too low. A lot of 60Hz
equipment won't work (well) on 50 Hz unless the transformer is specifically designed
for that. However, how that applies to a transformer that is already intended to saturate
I don't know. I do suspect it is still an issue since it will likely saturate sooner
than expected, for some definition of sooner.
Will
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up.
Remember that the resonant frequency goes as 1/sqrt(L*C). So doesn't
that mean you want a capacitor of 1.2^2 times the original value?
-tony 
12:42 p.m.
On Wed, 2025-07-16 at 11:42 -0400, Will Cooke via cctalk wrote:
The Computer History Museum in Sunnyvale, CA has a working IBM 1401
computer from Germany. It has ferroresonant power supplies. They bought
a converter to supply 50 Hz power because they were certain it wouldn't
work at 60 Hz. And it has motors in the card reader, card punch,
printer, and tape drives, that would all run at the wrong speed using
60 Hz power.
At first they had an antique and unreliable motor-generator. Now they
have a switching supply that IBM provided (maybe donated).
On 07/16/2025 11:10 AM EDT Tony Duell via
cctalk
<cctalk(a)classiccmp.org> wrote:
On Wed, Jul 16, 2025 at 4:00 PM Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
One thing to keep in mind is that the transformer may not be designed
to work at 50 Hz. A "normal" transformer will saturate if the
frequency is too low. A lot of 60Hz equipment won't work (well) on
50 Hz unless the transformer is specifically designed for that.
However, how that applies to a transformer that is already intended
to saturate I don't know. I do suspect it is still an issue since it
will likely saturate sooner than expected, for some definition of
sooner.
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up.
Remember that the resonant frequency goes as 1/sqrt(L*C). So
doesn't
that mean you want a capacitor of 1.2^2 times the original value?
-tony
Will
1:16 p.m.
I was thinking along the lines of Ben's and John's comment, but using a 12v
PS, a battery, and a 50 HZ UPS/ BBU.
Old school, low tech, but it works
Paul
On Wed, Jul 16, 2025 at 2:50 PM Van Snyder via cctalk <cctalk(a)classiccmp.org>
wrote:
On Wed, 2025-07-16 at 11:42 -0400, Will Cooke via
cctalk wrote:
The Computer History Museum in Sunnyvale, CA has a working IBM 1401
computer from Germany. It has ferroresonant power supplies. They bought
a converter to supply 50 Hz power because they were certain it wouldn't
work at 60 Hz. And it has motors in the card reader, card punch,
printer, and tape drives, that would all run at the wrong speed using
60 Hz power.
At first they had an antique and unreliable motor-generator. Now they
have a switching supply that IBM provided (maybe donated).
On 07/16/2025 11:10 AM EDT Tony Duell via
cctalk
<cctalk(a)classiccmp.org> wrote:
On Wed, Jul 16, 2025 at 4:00 PM Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
One thing to keep in mind is that the transformer may not be designed
to work at 50 Hz. A "normal" transformer will saturate if the
frequency is too low. A lot of 60Hz equipment won't work (well) on
50 Hz unless the transformer is specifically designed for that.
However, how that applies to a transformer that is already intended
to saturate I don't know. I do suspect it is still an issue since it
will likely saturate sooner than expected, for some definition of
sooner.
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up.
Remember that the resonant frequency goes as 1/sqrt(L*C). So
doesn't
that mean you want a capacitor of 1.2^2 times the original value?
-tony
Will
17 Jul
17 Jul
6:45 a.m.
On Jul 16, 2025, at 3:42 PM, Van Snyder via cctalk
<cctalk(a)classiccmp.org> wrote:
I wonder about "they were certain it would not work". That should be a question
of fact, not belief.
As for the motors, that's an obvious issue (if they are induction motors rather than
universal motors). The modern solution is a VFC -- variable frequency motor controller.
Those are pretty cheap and work great with motors. I've heard that they are not so
good with power supply transformers, not sure if that has been experimentally confirmed.
For power supply transformers, 50 vs. 60 Hz is unlikely to matter. People with CDC
mainframes that want 400 Hz power do need a solution, with motor-generators as the
traditional answer. I wonder if a VFC would work for that, perhaps with post-VFC
filtering to turn the waveform into something closer to a sine wave.
paul
...
The Computer History Museum in Sunnyvale, CA has a working IBM 1401
computer from Germany. It has ferroresonant power supplies. They bought
a converter to supply 50 Hz power because they were certain it wouldn't
work at 60 Hz. And it has motors in the card reader, card punch,
printer, and tape drives, that would all run at the wrong speed using
60 Hz power.
7:41 a.m.
On 7/17/25 08:45, Paul Koning via cctalk wrote:
VFDs (variable frequency drives) produce ~340 V PWM "square"
waves, Given enough inductance in the motor windings, this
causes roughly sinusoidal currents. But, feed this into a
transformer, and you will get high frequency spikes. Now,
MAYBE, due to the way a "Sola" transformer works, it might
smooth out the square waves, but it is real hard to predict
what will happen. Also, VFDs are usually designed for
balanced 3-phase loads.
Jon
On Jul 16, 2025, at 3:42 PM, Van Snyder via
cctalk <cctalk(a)classiccmp.org> wrote:
I wonder about "they were certain it would not work". That
should be a question of fact, not belief.
As for the motors, that's an obvious issue (if they are induction motors rather than
universal motors). The modern solution is a VFC -- variable frequency motor controller.
Those are pretty cheap and work great with motors. I've heard that they are not so
good with power supply transformers, not sure if that has been experimentally confirmed.
For power supply transformers, 50 vs. 60 Hz is unlikely to matter. People with CDC
mainframes that want 400 Hz power do need a solution, with motor-generators as the
traditional answer. I wonder if a VFC would work for that, perhaps with post-VFC
filtering to turn the waveform into something closer to a sine wave.
...
The Computer History Museum in
Sunnyvale, CA has a working IBM 1401
computer from Germany. It has ferroresonant power supplies. They bought
a converter to supply 50 Hz power because they were certain it wouldn't
work at 60 Hz. And it has motors in the card reader, card punch,
printer, and tape drives, that would all run at the wrong speed using
60 Hz power. 
8:37 a.m.
https://kikusuiamerica.com/products-index/ac/pcr-ma/ will oblige
I have a PCR500MA : good for 400 Hz synchro work, also 40/50/60 Hz induction motors - IIRC
the wfm is much cleaner than the usu 12V -> ~ static inverter. The pricing reflects
this.
3 phase (Op) VFD are wonderful for driving 3 phase machine tool motors, and can go slowly
ie low frequency; eg https://uk.rs-online.com/web/p/inverter-drives/7858522
I have a 1 phase (Op) VFD on the shelf, boat anchor, cf PCR500 - wfm is ugly
HtH B
-----Original Message-----
From: Jon Elson via cctalk [mailto:cctalk@classiccmp.org]
Sent: 17 July 2025 14:42
To: Paul Koning via cctalk <cctalk(a)classiccmp.org>
Cc: Jon Elson <elson(a)pico-systems.com>
Subject: [cctalk] Re: Ferroresonant transformer mystery
On 7/17/25 08:45, Paul Koning via cctalk wrote:
VFDs (variable frequency drives) produce ~340 V PWM "square"
waves, Given enough inductance in the motor windings, this causes roughly sinusoidal
currents. But, feed this into a transformer, and you will get high frequency spikes.
Now, MAYBE, due to the way a "Sola" transformer works, it might smooth out the
square waves, but it is real hard to predict what will happen. Also, VFDs are usually
designed for balanced 3-phase loads.
Jon
On Jul 16, 2025, at 3:42 PM, Van Snyder via
cctalk <cctalk(a)classiccmp.org> wrote:
I wonder about "they were certain it would not work". That should be a
question of fact, not belief.
As for the motors, that's an obvious issue (if they are induction motors rather than
universal motors). The modern solution is a VFC -- variable frequency motor controller.
Those are pretty cheap and work great with motors. I've heard that they are not so
good with power supply transformers, not sure if that has been experimentally confirmed.
For power supply transformers, 50 vs. 60 Hz is unlikely to matter. People with CDC
mainframes that want 400 Hz power do need a solution, with motor-generators as the
traditional answer. I wonder if a VFC would work for that, perhaps with post-VFC
filtering to turn the waveform into something closer to a sine wave.
...
The Computer History Museum in Sunnyvale, CA has a working IBM 1401
computer from Germany. It has ferroresonant power supplies. They
bought a converter to supply 50 Hz power because they were certain it
wouldn't work at 60 Hz. And it has motors in the card reader, card
punch, printer, and tape drives, that would all run at the wrong
speed using
60 Hz power.
9:01 a.m.
On Jul 17, 2025, at 10:41 AM, Jon Elson via cctalk
<cctalk(a)classiccmp.org> wrote:
On 7/17/25 08:45, Paul Koning via cctalk wrote:
VFDs (variable frequency drives) produce ~340 V PWM "square" waves, Given
enough inductance in the motor windings, this causes roughly sinusoidal currents. But,
feed this into a transformer, and you will get high frequency spikes. Now, MAYBE, due to
the way a "Sola" transformer works, it might smooth out the square waves, but it
is real hard to predict what will happen. Also, VFDs are usually designed for balanced
3-phase loads.
Yes, the question is whether the transformer could be made sufficiently happy. As for
3-phase: the CDC mainframes (6000 series) in fact use 3 phase 400 Hz power feeding their
power supplies -- you're dealing with 3-phase transformers and 6-diode rectifiers
behind them.
paul
On Jul 16, 2025, at 3:42 PM, Van Snyder via
cctalk <cctalk(a)classiccmp.org> wrote:
I wonder about "they were certain it would not work". That
should be a question of fact, not belief.
As for the motors, that's an obvious issue (if they are induction motors rather than
universal motors). The modern solution is a VFC -- variable frequency motor controller.
Those are pretty cheap and work great with motors. I've heard that they are not so
good with power supply transformers, not sure if that has been experimentally confirmed.
For power supply transformers, 50 vs. 60 Hz is unlikely to matter. People with CDC
mainframes that want 400 Hz power do need a solution, with motor-generators as the
traditional answer. I wonder if a VFC would work for that, perhaps with post-VFC
filtering to turn the waveform into something closer to a sine wave.
...
The Computer History Museum in
Sunnyvale, CA has a working IBM 1401
computer from Germany. It has ferroresonant power supplies. They bought
a converter to supply 50 Hz power because they were certain it wouldn't
work at 60 Hz. And it has motors in the card reader, card punch,
printer, and tape drives, that would all run at the wrong speed using
60 Hz power. 
9:55 a.m.
Paul Koning via cctalk wrote:
> I wonder about "they were certain it would
not work". That should be a question of fact, not belief.
-snip-
For power supply transformers, 50 vs. 60 Hz is
unlikely to matter. People with CDC mainframes that want 400 Hz power do need a solution,
with motor-generators as the traditional answer. I wonder if a VFC would work for that,
perhaps with post-VFC filtering to turn the waveform into something closer to a sine
wave.
paul
Operating a 60Hz transformer at 50Hz will require a higher magnetizing
current (and flux density) in order to maintain the same operating
voltage. This is very likely to send the transformer into a B vs H
region with reduced permeability, thus creating a much more distorted
magnetizing current waveform with sizable peaks. This will in turn
increase primary copper losses, and overheating can happen easily,
depending on the transformer specifics. For power distribution and
transmission transformers, this is definitely a no-no. There is a
reason that 50Hz utility transformers have roughly 20% more iron and
copper in them than 60Hz ones.
Carlos.
11:55 a.m.
On Jul 17, 2025, at 12:55 PM, Carlos E Murillo-Sanchez
via cctalk <cctalk(a)classiccmp.org> wrote:
Paul Koning via cctalk wrote:
Yes, I can see that for utility transformers, which are carefully optimized for the task.
But I would think that power supply transformers are likely to have larger margins so they
aren't operated so close to the limits.
paul
> I wonder about "they were certain it
would not work". That should be a question of fact, not belief.
-snip-
For power supply transformers, 50 vs. 60 Hz is
unlikely to matter. People with CDC mainframes that want 400 Hz power do need a solution,
with motor-generators as the traditional answer. I wonder if a VFC would work for that,
perhaps with post-VFC filtering to turn the waveform into something closer to a sine
wave.
paul
Operating a 60Hz transformer at 50Hz will require a higher magnetizing
current (and flux density) in order to maintain the same operating voltage. This is very
likely to send the transformer into a B vs H region with reduced permeability, thus
creating a much more distorted magnetizing current waveform with sizable peaks. This will
in turn increase primary copper losses, and overheating can happen easily, depending on
the transformer specifics. For power distribution and transmission transformers, this is
definitely a no-no. There is a reason that 50Hz utility transformers have roughly 20%
more iron and copper in them than 60Hz ones.
12:30 p.m.
Yes, I can see that for utility transformers, which are carefully optimized for the task.
But I would think that power supply transformers are likely to have larger margins so
they aren't operated so close to the limits.
paul
Here is a list of current production models of power transformers. Notice it has separate
models for 50/60 Hz and 60 Hz.
Many years ago, I bought an Epson PX-8 just prior to moving to Germany from the US. Using
a 220-110 transformer, it would take about 2 days to charge the battery, more than twice
as long as in the US. But the wall-wart transformer got very warm, much warmer than in
the US. Upon returning to the US with our 60 Hz power, all was well. The battery charged
fully in less than a day and the transformer only got slightly warm.
Just a data point.
Will
12:31 p.m.
Well, I suppose I should add the link to the data I mentioned.
https://www.hammfg.com/electronics/transformers/power/166
16 Jul
16 Jul
2:06 p.m.
Hi Jon
I restored a PDP-8 straight 8 that had a similar PSU. I am here in New Zealand with 230v
50Hz. I used a step down transformer 230v -> 110v but of course retained the 50hz.
It worked just fine. I did replace the capacitor in question however. They contain PCB’s
so use gloves and dispose with care.
Kind regards
Brendan
--------------//----------------
brendan@mcneill.co.nz<mailto:brendan@mcneill.co.nz>
+64 21 881 883
From: Jon Elson via cctalk <cctalk(a)classiccmp.org>
Date: Thursday, 17 July 2025 at 02:50
To: Tom Hunter via cctalk <cctalk(a)classiccmp.org>
Cc: Jon Elson <elson(a)pico-systems.com>
Subject: [cctalk] Re: Ferroresonant transformer mystery
On 7/16/25 09:13, Tom Hunter via cctalk wrote:
Recently I got a nice and complete PDP-8/s from the
US. The power supply
uses a ferroresonant transformer which in addition to the standard primary
and secondary windings has a separate 2.3H winding connected in series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The transformer's
secondary side and the resonant circuit are operated in saturation. There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to the
input frequency of 60Hz. This is quite a nice if not elegant design for the
period in question, but maybe not the most efficient.
As I live in Australia I get 240VAC and 50Hz as opposed to the US 115VAC
and 60Hz.
I can easily convert our 240VAC to 115VAC with a step-down transformer, but
cannot easily supply 115VAC at 60Hz. So I was considering using a step-down
transformer to get the 115VAC, but modify the resonant "tank circuit" for
50Hz.
Unfortunately there is some magic I don't understand. The resonant
frequency of a LC circuit with L=2.3H and C=2uF is about 75Hz not the
expected 60Hz.
Otherwise I could just solve the standard LC resonant circuit formula for C
and plug in 50Hz and 2.3H to get the required C.
Obviously ferroresonant transformers are more complex than this former
software engineer can grasp. Could any experienced EE with relevant
transformer knowledge please chime in and help me understand how to
redimension the tank circuit to use 50Hz instead of the original 60Hz input.
Obviously I could replace the entire power supply with two modern switch
mode supplies to create the two rails, but it would be really nice to keep
the original supply and just reversibly adapt it for 50Hz.
That L is probably different when the transformer is excited
at full mains voltage and near saturation. Of course, this
seems like it will reduce L and therefore drive the
resonance higher!
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up. Then, measure the output
voltage, and if it is within range just use it like that.
If you want to get fancy, put it on a Variac and sweep the
input voltage. You will note a reverse slope, as input
voltage rises through the range, output voltage will decline.
Jon
3:09 p.m.
I found this,
https://www.solidstatecontrolsinc.com/-/media/ameteksolidstatecontrols/docu…
which says in section 1.8 on page 8,
1.8 FREQUENCY SENSITIVITY
Referring to Figure 1.3, we see that the slope of the straight line
varies inversely with the square of the input frequency. As the frequency
increases, the slope of the straight line decreases and point C moves up
to a higher value of flux density. The output voltage of the regulator
is, therefore, sensitive to frequency variations, the relation being a
1.4% change in output voltage for a 1% change in input frequency
Chris
On Wednesday (07/16/2025 at 09:06PM +0000), Brendan McNeill via cctalk wrote:
Hi Jon
I restored a PDP-8 straight 8 that had a similar PSU. I am here in New Zealand with 230v
50Hz. I used a step down transformer 230v -> 110v but of course retained the 50hz.
It worked just fine. I did replace the capacitor in question however. They contain PCB’s
so use gloves and dispose with care.
Kind regards
Brendan
--------------//----------------
brendan@mcneill.co.nz<mailto:brendan@mcneill.co.nz>
+64 21 881 883
From: Jon Elson via cctalk <cctalk(a)classiccmp.org>
Date: Thursday, 17 July 2025 at 02:50
To: Tom Hunter via cctalk <cctalk(a)classiccmp.org>
Cc: Jon Elson <elson(a)pico-systems.com>
Subject: [cctalk] Re: Ferroresonant transformer mystery
On 7/16/25 09:13, Tom Hunter via cctalk wrote:
--
Chris Elmquist
Recently I got a nice and complete PDP-8/s from
the US. The power supply
uses a ferroresonant transformer which in addition to the standard primary
and secondary windings has a separate 2.3H winding connected in series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The transformer's
secondary side and the resonant circuit are operated in saturation. There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to the
input frequency of 60Hz. This is quite a nice if not elegant design for the
period in question, but maybe not the most efficient.
As I live in Australia I get 240VAC and 50Hz as opposed to the US 115VAC
and 60Hz.
I can easily convert our 240VAC to 115VAC with a step-down transformer, but
cannot easily supply 115VAC at 60Hz. So I was considering using a step-down
transformer to get the 115VAC, but modify the resonant "tank circuit" for
50Hz.
Unfortunately there is some magic I don't understand. The resonant
frequency of a LC circuit with L=2.3H and C=2uF is about 75Hz not the
expected 60Hz.
Otherwise I could just solve the standard LC resonant circuit formula for C
and plug in 50Hz and 2.3H to get the required C.
Obviously ferroresonant transformers are more complex than this former
software engineer can grasp. Could any experienced EE with relevant
transformer knowledge please chime in and help me understand how to
redimension the tank circuit to use 50Hz instead of the original 60Hz input.
Obviously I could replace the entire power supply with two modern switch
mode supplies to create the two rails, but it would be really nice to keep
the original supply and just reversibly adapt it for 50Hz.
That L is probably different when the transformer is excited
at full mains voltage and near saturation. Of course, this
seems like it will reduce L and therefore drive the
resonance higher!
What I might do is make up a cap bank that is 1.2 X larger
than the 2 uF and power it up. Then, measure the output
voltage, and if it is within range just use it like that.
If you want to get fancy, put it on a Variac and sweep the
input voltage. You will note a reverse slope, as input
voltage rises through the range, output voltage will decline.
Jon 
8:57 a.m.
On 2025-07-16 8:13 a.m., Tom Hunter via cctalk wrote:
Recently I got a nice and complete PDP-8/s from the
US. The power supply
uses a ferroresonant transformer which in addition to the standard primary
and secondary windings has a separate 2.3H winding connected in series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The transformer's
secondary side and the resonant circuit are operated in saturation. There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to the
input frequency of 60Hz. This is quite a nice if not elegant design for the
period in question, but maybe not the most efficient.
A 50 hz to 12? volt DC and 12? volt DC to 60 hz converter might be a
better option if you have moving parts, plus you have safe power supply.
Have you checked for a UK pdp 8 power supply schematic?
Ben.
11:05 a.m.
On 2025-07-16 8:57 a.m., ben via cctalk wrote:
On 2025-07-16 8:13 a.m., Tom Hunter via cctalk wrote:
Exactly what I do to change motor speeds on my jukeboxes - convert the
line to 12VDC, then use an Inverter for automobiles, etc, that puts out
120 @ 60 Hz at the current required. I then modify the converter output
frequency if I want to speed up or slow down the synchronous motor speed...
John :-#)#
--
John's Jukes Ltd.
7 - 3979 Marine Way, Burnaby, BC, Canada V5J 5E3
Call (604)872-5757 (Pinballs, Jukes, Video Games)
flippers.com
"Old pinballers never die, they just flip out"
Recently I got a nice and complete PDP-8/s from
the US. The power supply
uses a ferroresonant transformer which in addition to the standard
primary
and secondary windings has a separate 2.3H winding connected in
series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The
transformer's
secondary side and the resonant circuit are operated in saturation.
There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to
the
input frequency of 60Hz. This is quite a nice if not elegant design
for the
period in question, but maybe not the most efficient.
A 50 hz to 12? volt DC and 12? volt DC to 60 hz converter might be a
better option if you have moving parts, plus you have safe power supply.
Have you checked for a UK pdp 8 power supply schematic?
Ben.
31 Jul
31 Jul
11:11 a.m.
On Wed, 16 Jul 2025, Tom Hunter via cctalk wrote:
Recently I got a nice and complete PDP-8/s from the
US. The power supply
uses a ferroresonant transformer which in addition to the standard primary
and secondary windings has a separate 2.3H winding connected in series to a
2uF 660VAC capacitor forming a resonant "tank" circuit. The transformer's
secondary side and the resonant circuit are operated in saturation. There
is a magnetic shunt to prevent the primary side going into saturation as
well. It accepts a wide input voltage range, but is very sensitive to the
input frequency of 60Hz. This is quite a nice if not elegant design for the
period in question, but maybe not the most efficient.
As I live in Australia I get 240VAC and 50Hz as opposed to the US 115VAC
and 60Hz.
Late input and FWIW. As taking care of the old pieces of equipment is
enough of a challenge by itself I try to reduce the distraction by using
modern and off-the-shelf lab support infrastructure components.
In this case I would just go for a proper power converter such as one
listed here: <https://www.amazon.com/dp/B09KVG9N71/> (not endorsing the
manufacturer, the product, or the web store in any way), having checked
that the power budget is sufficient of course, as this haven't been stated
for the machine in question. Also to keep the original system intact.
Maciej
11
days inactive
26
days old
19 comments
14 participants
participants (14)
-
ben -
Brendan McNeill -
Carlos E Murillo-Sanchez -
Chris Elmquist -
John Robertson -
Jon Elson -
Maciej W. Rozycki -
Martin Bishop -
Paul Anderson -
Paul Koning -
Tom Hunter -
Tony Duell -
Van Snyder -
wrcooke＠wrcooke.net