Paper caps all replaced, and a couple out-of-tolerance cathode resistors
swapped out for 1 watt carbon film. The 6F6 output stage cathode bypass
electrolytic was also replaced; it wasn't in the can as specified in the
schematic but was a "Dandee" brand 10uF axial lead type air-wired inside
the chassis.

The paper caps were a hodge-podge of different brands - Mallory,
Aerovox, Sprague, Tobe Detustchmann. The can electrolytic is an 8uF/8uF
unit instead of the specified 6-6-6. Some of the paper caps weren't the
correct values e.g. the 0.25uF paper cap was actually a (very large)
0.5uF unit. I'm guessing the last time this radio was serviced might
have been during WW2 and perhaps wartime rationing had something to do
with it - repair shop made use of whatever they had on hand.

Maybe surprisingly both sections of the can electrolytic reformed OK
running a current thru it via stepping up the voltage ranges while on
the "leakage" setting of my Heathkit IT-11. It took about a minute for
the eye to open at the rated max voltage of 450. I measure a "power
factor" of around 3-4% on both sections at that voltage. Don't know if
it would be good to leave that as is but I'm using it in-place for testing.

The 5Z4 and 6F6 + power supply circuit seems to be working fine, with
the set powered up the voltage on the first capacitor section and the
second after the field coil choke look approximately correct.
Injecting1kHz into the 6F6 grid and into the grid of the 6J7 detector I
get a strong output through the speaker with no hum.

The problem now is that the plate and screen voltages on the 6A8 and 6K7
are way off; the service manual specifies 195 and 210 for plate voltages
respectively but I'm actually getting more like 260 and 275. The screen
voltages are specified as being 70 and 90 respectively but I'm looking
at about 130 for both of them. Pin 5 of the 6A8 looks correct at about
-10.5 volts. Cathode resistors and bypasses were replaced on both of
them but the cathode voltages are high, e.g. with the AM broadcast band
selected, volume all the way up I'm reading about 4.8 volts on the
cathode of the 6A8 instead of 3.

With a 3 meter longwire antenna I'm getting one local strong AM station
(50kW about 8 miles down the road) at night but the AM band is otherwise
silent. I'm guessing at this point one or both of the 6A8/6K7 has poor
emission? I don't own a vintage tube tester but I have both HV and low
voltage bench supplies available, any suggestion on how to rig up a
quick emissions test?

On 11/22/2017 09:01 AM, bitrex wrote:
Paper caps all replaced, and a couple out-of-tolerance cathode resistors
swapped out for 1 watt carbon film. The 6F6 output stage cathode bypass
electrolytic was also replaced; it wasn't in the can as specified in the
schematic but was a "Dandee" brand 10uF axial lead type air-wired inside
the chassis.

The paper caps were a hodge-podge of different brands - Mallory,
Aerovox, Sprague, Tobe Detustchmann. The can electrolytic is an 8uF/8uF
unit instead of the specified 6-6-6. Some of the paper caps weren't the
correct values e.g. the 0.25uF paper cap was actually a (very large)
0.5uF unit. I'm guessing the last time this radio was serviced might
have been during WW2 and perhaps wartime rationing had something to do
with it - repair shop made use of whatever they had on hand.

Maybe surprisingly both sections of the can electrolytic reformed OK
running a current thru it via stepping up the voltage ranges while on
the "leakage" setting of my Heathkit IT-11. It took about a minute for
the eye to open at the rated max voltage of 450. I measure a "power
factor" of around 3-4% on both sections at that voltage. Don't know if
it would be good to leave that as is but I'm using it in-place for testing.

The 5Z4 and 6F6 + power supply circuit seems to be working fine, with
the set powered up the voltage on the first capacitor section and the
second after the field coil choke look approximately correct.
Injecting1kHz into the 6F6 grid and into the grid of the 6J7 detector I
get a strong output through the speaker with no hum.

The problem now is that the plate and screen voltages on the 6A8 and 6K7
are way off; the service manual specifies 195 and 210 for plate voltages
respectively but I'm actually getting more like 260 and 275. The screen
voltages are specified as being 70 and 90 respectively but I'm looking
at about 130 for both of them. Pin 5 of the 6A8 looks correct at about
-10.5 volts. Cathode resistors and bypasses were replaced on both of
them but the cathode voltages are high, e.g. with the AM broadcast band
selected, volume all the way up I'm reading about 4.8 volts on the
cathode of the 6A8 instead of 3.

With a 3 meter longwire antenna I'm getting one local strong AM station
(50kW about 8 miles down the road) at night but the AM band is otherwise
silent. I'm guessing at this point one or both of the 6A8/6K7 has poor
emission? I don't own a vintage tube tester but I have both HV and low
voltage bench supplies available, any suggestion on how to rig up a
quick emissions test?



Schematic: http://www.nostalgiaair.org/pagesbymodel/542/M0009542.pdf

On 11/22/2017 09:01 AM, bitrex wrote:

The problem now is that the plate and screen voltages on the 6A8 and 6K7
are way off; the service manual specifies 195 and 210 for plate voltages
respectively but I'm actually getting more like 260 and 275. The screen
voltages are specified as being 70 and 90 respectively but I'm looking
at about 130 for both of them. Pin 5 of the 6A8 looks correct at about
-10.5 volts. Cathode resistors and bypasses were replaced on both of
them but the cathode voltages are high, e.g. with the AM broadcast band
selected, volume all the way up I'm reading about 4.8 volts on the
cathode of the 6A8 instead of 3.

BTW I also replaced the screen resistor feeding both screen circuits as
well as the screen bypass cap. No change.

What is your wallplate voltage?

110:195 = XXX/260 where XXX = 147 VAC.
110:210 = XXX/275 where XXX = 144 VAC.
110:3 = XXX/4.8 where XXX = 179 VAC

These results are reasonably self-consistent as what comes to a volume-control can have many variations. Note that increasing capacitance in many cases will also increase B+ and other down-line voltages, and generally *SHOULD NOT BE DONE*, especially in radios with field-coil speakers such as that one.

So, hazarding a guess at long distance, you have a combination of high wallplate voltage, excessive capacitance and measurement variance.


This voltage of 14X seems extreme, but given that these sets were measured with VTVMs back in the day, and VTVMs load very differently than a modern VOM. And the vintage unit was not "true RMS" and so could measure low. Do you have an old analog meter? Or, perhaps a VTVM?

I have long-since learned to distrust factory-schematic voltages. Generally.. I will bring a radio up to 110V on the Variac and measure secondary voltages off the transformer (unloaded, then loaded) to see what is what. If I get consistent discrepancies, I attribute it to measurement variances. If I get inconsistent discrepancies - more than 15% apart - I will look for a bad transformer winding or some other problem of that nature.

Did you replace the line-filter cap on the line-cord? Do so with a type Y safety capacitor. That will help on reception.

I have no idea where you are, but if you are anywhere near southeastern PA, USA, I would be glad to test your tubes. I keep both a little Simpson emissions tester and a Hickok 539B for the heavy-duty stuff.

Otherwise, try swapping tubes with another radio. 6A8s are notoriously prone to failure, and glass ones are getting quite scarce. More so than even a 1L6 in my experience.

Peter Wieck
Melrose Park, PA

On 11/22/2017 09:54 AM, wrote:
What is your wallplate voltage?

110:195 = XXX/260 where XXX = 147 VAC.
110:210 = XXX/275 where XXX = 144 VAC.
110:3 = XXX/4.8 where XXX = 179 VAC

These results are reasonably self-consistent as what comes to a volume-control can have many variations. Note that increasing capacitance in many cases will also increase B+ and other down-line voltages, and generally *SHOULD NOT BE DONE*, especially in radios with field-coil speakers such as that one.

Using my DMM on the "AC" setting, out of the wall it's reading 124 VAC.

So, hazarding a guess at long distance, you have a combination of high wallplate voltage, excessive capacitance and measurement variance.


This voltage of 14X seems extreme, but given that these sets were measured with VTVMs back in the day, and VTVMs load very differently than a modern VOM. And the vintage unit was not "true RMS" and so could measure low. Do you have an old analog meter? Or, perhaps a VTVM?

I have long-since learned to distrust factory-schematic voltages. Generally. I will bring a radio up to 110V on the Variac and measure secondary voltages off the transformer (unloaded, then loaded) to see what is what. If I get consistent discrepancies, I attribute it to measurement variances. If I get inconsistent discrepancies - more than 15% apart - I will look for a bad transformer winding or some other problem of that nature.

Did you replace the line-filter cap on the line-cord? Do so with a type Y safety capacitor. That will help on reception.

I have no idea where you are, but if you are anywhere near southeastern PA, USA, I would be glad to test your tubes. I keep both a little Simpson emissions tester and a Hickok 539B for the heavy-duty stuff.

Thanks for the offer! I'm currently in Providence, RI most of the time
so a little far, unfortunately...

Otherwise, try swapping tubes with another radio. 6A8s are notoriously prone to failure, and glass ones are getting quite scarce. More so than even a 1L6 in my experience.

Peter Wieck
Melrose Park, PA


Will do. Yeah, gosh, NOS 6A8Gs are expensive! The schematic specifies
either the metal or "G" shape enclosure, I guess the GT doesn't have the
correct pinout? In any case, all the tubes currently fitted in the set
are the metal enclosure type.

On 11/22/2017 09:54 AM, wrote:
What is your wallplate voltage?

110:195 = XXX/260 where XXX = 147 VAC.
110:210 = XXX/275 where XXX = 144 VAC.
110:3 = XXX/4.8 where XXX = 179 VAC

These results are reasonably self-consistent as what comes to a volume-control can have many variations. Note that increasing capacitance in many cases will also increase B+ and other down-line voltages, and generally *SHOULD NOT BE DONE*, especially in radios with field-coil speakers such as that one.

Seems extreme for a 6uF - 8uF swap, but I'm not experienced enough with
tube PSUs to know for sure. Weird that the cathode resistor voltage is
too high, I'd think that if emission were the problem it would be too
low. There's definitely not enough voltage drop across the screen
resistor. I'm going to measure every voltage on every pin again and post
it to make sure I haven't bungled a measurement, somewhere.

So, hazarding a guess at long distance, you have a combination of high wallplate voltage, excessive capacitance and measurement variance.


This voltage of 14X seems extreme, but given that these sets were measured with VTVMs back in the day, and VTVMs load very differently than a modern VOM. And the vintage unit was not "true RMS" and so could measure low. Do you have an old analog meter? Or, perhaps a VTVM?

Unfortunately not, all my meters are DMMs. Unlike a tube tester that's
probably a tool I can get my hands on readily though

I have long-since learned to distrust factory-schematic voltages. Generally. I will bring a radio up to 110V on the Variac and measure secondary voltages off the transformer (unloaded, then loaded) to see what is what. If I get consistent discrepancies, I attribute it to measurement variances. If I get inconsistent discrepancies - more than 15% apart - I will look for a bad transformer winding or some other problem of that nature.

Did you replace the line-filter cap on the line-cord? Do so with a type Y safety capacitor. That will help on reception.

Ok, will do, it's currently just a 630V regular film type.

I have no idea where you are, but if you are anywhere near southeastern PA, USA, I would be glad to test your tubes. I keep both a little Simpson emissions tester and a Hickok 539B for the heavy-duty stuff.

Otherwise, try swapping tubes with another radio. 6A8s are notoriously prone to failure, and glass ones are getting quite scarce. More so than even a 1L6 in my experience.

Peter Wieck
Melrose Park, PA

On Wednesday, November 22, 2017 at 10:22:52 AM UTC-5, bitrex wrote:

Will do. Yeah, gosh, NOS 6A8Gs are expensive! The schematic specifies
either the metal or "G" shape enclosure, I guess the GT doesn't have the
correct pinout? In any case, all the tubes currently fitted in the set
are the metal enclosure type.

The GT is fine, but it is a short tube and often the grid-cap lead will not reach the top of the tube over the shield - required for a glass-type tube. If you have no shield, stick with the metal tubes. \

It may also need an alignment, by the way. Are you equipped to do that? It is certainly simple enough - with the right tools.

Peter Wieck
Melrose Park, PA

On 11/22/2017 09:54 AM, wrote:
What is your wallplate voltage?

110:195 = XXX/260 where XXX = 147 VAC.
110:210 = XXX/275 where XXX = 144 VAC.
110:3 = XXX/4.8 where XXX = 179 VAC

These results are reasonably self-consistent as what comes to a volume-control can have many variations. Note that increasing capacitance in many cases will also increase B+ and other down-line voltages, and generally *SHOULD NOT BE DONE*, especially in radios with field-coil speakers such as that one.

So, hazarding a guess at long distance, you have a combination of high wallplate voltage, excessive capacitance and measurement variance.


This voltage of 14X seems extreme, but given that these sets were measured with VTVMs back in the day, and VTVMs load very differently than a modern VOM. And the vintage unit was not "true RMS" and so could measure low. Do you have an old analog meter? Or, perhaps a VTVM?

I have long-since learned to distrust factory-schematic voltages. Generally. I will bring a radio up to 110V on the Variac and measure secondary voltages off the transformer (unloaded, then loaded) to see what is what. If I get consistent discrepancies, I attribute it to measurement variances. If I get inconsistent discrepancies - more than 15% apart - I will look for a bad transformer winding or some other problem of that nature.

Did you replace the line-filter cap on the line-cord? Do so with a type Y safety capacitor. That will help on reception.

I have no idea where you are, but if you are anywhere near southeastern PA, USA, I would be glad to test your tubes. I keep both a little Simpson emissions tester and a Hickok 539B for the heavy-duty stuff.

Otherwise, try swapping tubes with another radio. 6A8s are notoriously prone to failure, and glass ones are getting quite scarce. More so than even a 1L6 in my experience.

Peter Wieck
Melrose Park, PA


Updated voltage measurements, broadcast band, volume control fully on,
wallplug voltage 124VAC. All measurements taken ~30 sc

6F6:

Pin 1, enclosu 0
Pin 2, heater: 3.2 VAC
Pin 3, plate: 283
Pin 4, screen grid: 283
Pin 5, control grid: 0.8
Pin 6, NC
Pin 7, heater: 3.2 VAC
Pin 8, cathode: 18

6J7:

Pin 1, enclosu 0
Pin 2, heater 3.2VAC
Pin 3, plate: 7.7 ------ wut
Pin 4, screen grid: 18
Pin 5, suppressor: 1.8
Pin 6, NC
Pin 7, heater: 3.2VAC
Pin 8, cathode: 1.8

grid, top cap: 1mV

6K7:

Pin 1, enclosu 0
Pin 2, heater: 3.2VAC
Pin 3, plate: 284
Pin 4, screen: 127
Pin 5, suppressor: 3.5
Pin 6, NC
Pin 7, heater: 3.2Vac
Pin 8, cathode: 3.5

grid, top cap: 1mV

6A8:

Pin 1, enclosu 0
Pin 2, heater: 3.2VAC
Pin 3, plate: 267
Pin 4, screen: 100
Pin 5, grid 1: -10
Pin 6, grid 2: 235
Pin 7, heater: 3.2VAC
Pin 8, cathode: 4.3

grid, top cap: 2mV

On 11/22/2017 12:33 PM, bitrex wrote:

Updated voltage measurements, broadcast band, volume control fully on,
wallplug voltage 124VAC. All measurements taken ~30 sc

30 seconds after power on, rather

You are getting 6.4V across the filament - which makes sense at 124V at the wallplate.

I expect that your transformer is good.
I expect that the elevated B+ is due, in part, to excess capacitance. 6 - 8 uf may not seem like much from 30,000 feet, but it is a 33% increase at ground level. And 6 - 10 uF is a 67% increase at ground level. And, of course, an 11+% increase in wallplate voltage will contribute significantly.

Do you have any replacement tubes in your inventory?

If not, contact me off-group. I probably have dozens in that line-up and I could send you some proven-good for testing. Also, can you do an alignment?

Peter Wieck
Melrose Park, PA

On 11/22/2017 01:04 PM, wrote:
You are getting 6.4V across the filament - which makes sense at 124V at the wallplate.

I expect that your transformer is good.
I expect that the elevated B+ is due, in part, to excess capacitance. 6 - 8 uf may not seem like much from 30,000 feet, but it is a 33% increase at ground level. And 6 - 10 uF is a 67% increase at ground level. And, of course, an 11+% increase in wallplate voltage will contribute significantly.

Right, that makes sense, but what about the low plate voltage on the
6J7? There's a small RC network there, I don't think I checked all those
for tolerance. I should do that.

Do you have any replacement tubes in your inventory?

If not, contact me off-group. I probably have dozens in that line-up and I could send you some proven-good for testing. Also, can you do an alignment?

Peter Wieck
Melrose Park, PA


That'd be great, I can contribute a nominal fee if you like because I'm
definitely not thrilled with the prices I'm seeing on even used variants
at the usual outlets! Yeah, I have an RF signal gen available to do an
alignment, though I've only read about the procedure and never actually
done it in practice, doesn't seem too intimidating.