I've been wondering what is the correct way to wire a 3B7 filament . I've wired mine in parallel with the filament positive on pins 1/8 taking the cathode connection / filament negative from pin 4 which sits on an LED for bias . Is this correct for a DC filament supply for this type ?
316a
316a
Attachments
I have no comprehensive data for this tube, but it appears to be a loctal equivalent of the 3A5. This is a VHF portable transmitter type.
For audio use it will require negative bias on the grids.
With these tubes, datasheets assume that the filament cente tap is considered the refrence pin and for parallel operation of heaters, pins 1 & 8 should be positive wrt pin 4.
However, with a new tube it really doesn't matter - it will work equally well with pins 1 & 8 earthed and pin 4 positive. And since, unlike standard 1.4V battery receiving tubes it will require significant grid bias, it will work just as well with a negtive filament supply.
If the tube is a used tube, it can be a diffrent story. This is because, on any filament strand, the current due to emission partly cancels the heating current, and in the other half the currents add. This results in the more postive filament strand running at a slightly lower temperature and thus giving lower emission. So a used tube will have uneven "wear" on each filament strand. Thus which way round to connect the filaments may affect the available anode current under signal postive conditions.
Note that, in audio service, biasing with a cathode resistor cannot be used with this type of twin triode unless the tube is operated with both triodes in parallel or in push pull. Operation with the two triode sections in cascade will require a separate grid bias supply - without them you will have a multvibrator. In RF oscillator/power amplifier service that teh tube is designed for, grid bias would be derived from grid current charging grid capacitors.
Note that for audio power output service, a pair of 3V4 tetrodes will be more efficient. The filament power would be only slightly less, but the anode efficiency would be much greater.
For audio use it will require negative bias on the grids.
With these tubes, datasheets assume that the filament cente tap is considered the refrence pin and for parallel operation of heaters, pins 1 & 8 should be positive wrt pin 4.
However, with a new tube it really doesn't matter - it will work equally well with pins 1 & 8 earthed and pin 4 positive. And since, unlike standard 1.4V battery receiving tubes it will require significant grid bias, it will work just as well with a negtive filament supply.
If the tube is a used tube, it can be a diffrent story. This is because, on any filament strand, the current due to emission partly cancels the heating current, and in the other half the currents add. This results in the more postive filament strand running at a slightly lower temperature and thus giving lower emission. So a used tube will have uneven "wear" on each filament strand. Thus which way round to connect the filaments may affect the available anode current under signal postive conditions.
Note that, in audio service, biasing with a cathode resistor cannot be used with this type of twin triode unless the tube is operated with both triodes in parallel or in push pull. Operation with the two triode sections in cascade will require a separate grid bias supply - without them you will have a multvibrator. In RF oscillator/power amplifier service that teh tube is designed for, grid bias would be derived from grid current charging grid capacitors.
Note that for audio power output service, a pair of 3V4 tetrodes will be more efficient. The filament power would be only slightly less, but the anode efficiency would be much greater.
Last edited:
Yes, Parallel connexion is best, since the voltage gradient along the filament skews the effective bias. Lower voltage is better. The polarisation of the feed should not matter.
The original question was a sanity check 
Thanks to the responders . I have worked with DHT for years but find some of those battery filament types to be unpredictable , and even more unpredictable with reversed filament polarity . It turns out that although the 3B7 has low transconductance , it may require a grid stopper which I did not include . This may be due to my kludged layout but it appears a 100 ohm carbon comp in series with each grid cured the strange bias anomalies and the odd sound I was getting . What I now have has quiet 'cowbell' filament micophony for 5 minutes and impressive , but otherwise dark sound afterwards , definitely a bit of microphony euphonics coming through. 3B7 appears to require a backing off of filament volts to be useable ( running at 1.1V) and requires mechanical isolation , if you clap your hands next to the valve , there is a quiet 'clang' through the speakers . A nice sounding valve but quirky , I think a 6E5P may be a better option for the 4P1L output stage I'm using
316a
Thanks to the responders . I have worked with DHT for years but find some of those battery filament types to be unpredictable , and even more unpredictable with reversed filament polarity . It turns out that although the 3B7 has low transconductance , it may require a grid stopper which I did not include . This may be due to my kludged layout but it appears a 100 ohm carbon comp in series with each grid cured the strange bias anomalies and the odd sound I was getting . What I now have has quiet 'cowbell' filament micophony for 5 minutes and impressive , but otherwise dark sound afterwards , definitely a bit of microphony euphonics coming through. 3B7 appears to require a backing off of filament volts to be useable ( running at 1.1V) and requires mechanical isolation , if you clap your hands next to the valve , there is a quiet 'clang' through the speakers . A nice sounding valve but quirky , I think a 6E5P may be a better option for the 4P1L output stage I'm using316a
Don't back off the filament volts unless you are running it with anode current well below ratings. You'll get a short tube life otherwise.
If you need to reduce filament volts to get stability, it just means you layout requires a bit more grid resistance. 100 ohm is a very small amount for a grid stopper.
Bell-like microphonics is typical with the larger battery tubes, especially when used in low level stages. Some octal battery tubes were made with a mica "clamp" half way along the filament to bring the bell-like micophinics under control. Better quality commerical equipment mechanically isolated the sockets on leaf springs or, sometimes, a rubber grommet arrangment.
There are two main reasons for the characteristics being a bit changeable in old battery tubes:-
a) the differential wear out in each filament strand due to the emission current altering the filament temperature as I said above;
b) The filament scraping on the micas. As the filament heats up, it expands. For a filament strand 22mm long, the expansion is in the range 0.1 to 0.17 mm depending on the alloy used. This causes friction with the top mica, and slight temperature differences at the top end of the filament strand each time the tube is warmed up. Even on late design battery tubes such as the 1T4, if you monitor filament current with a 4.5 digit DVM while gradually altering filament volts, you can sometimes see small jumps in the current due to this reason. The early DHT tube designs, and almost all DHT tubes designed for AC, do not suffer from this problem as the filament is kept not in contact with a top mica.
If you need to reduce filament volts to get stability, it just means you layout requires a bit more grid resistance. 100 ohm is a very small amount for a grid stopper.
Bell-like microphonics is typical with the larger battery tubes, especially when used in low level stages. Some octal battery tubes were made with a mica "clamp" half way along the filament to bring the bell-like micophinics under control. Better quality commerical equipment mechanically isolated the sockets on leaf springs or, sometimes, a rubber grommet arrangment.
There are two main reasons for the characteristics being a bit changeable in old battery tubes:-
a) the differential wear out in each filament strand due to the emission current altering the filament temperature as I said above;
b) The filament scraping on the micas. As the filament heats up, it expands. For a filament strand 22mm long, the expansion is in the range 0.1 to 0.17 mm depending on the alloy used. This causes friction with the top mica, and slight temperature differences at the top end of the filament strand each time the tube is warmed up. Even on late design battery tubes such as the 1T4, if you monitor filament current with a 4.5 digit DVM while gradually altering filament volts, you can sometimes see small jumps in the current due to this reason. The early DHT tube designs, and almost all DHT tubes designed for AC, do not suffer from this problem as the filament is kept not in contact with a top mica.
Last edited:
I have never encountered issues running starved filaments , it is usually the best way of taming microphonics . 3B7 is a UHF type , it has low gm but those 100 ohm stoppers are definitely required . It's pretty pointless monitoring filament current when the filament is running constant current... After the valve warms up no pins and poings unless the valve is tapped , with the filament at 1.4V it is unuseable .
316a
Yes, over at the 4P1L line amp thread, most folks are running at 500-550mA (rating is approximately 650mA) for exactly this reason. In a line amp, the idle operating condition is not very stressful, but maybe the lifetime is lower. Andy Evans knows the answer to this, since he runs them without switching OFF. Somewhere, he reported what the lifetime came to. An adjustable-current filament regulator is certainly helpful for finding the optimum level.
The amp with 3B7 runs 4P1L as output stage . I'm running the 4P1L at 3.9V which is approximately the current suggested above . I'm getting away with it running 97dB sensitive speakers but wouldn't want to go higher !
316a
Starving the filament is ok PROVIDED the anode current is very much lower than rated anode current. In commercial equipment, running the filement or heater on reduced voltage is sometimes done to reduce noise, or to reduce drift in DC amplifiers, or to get better linearity in electronic voltmeters. But ALWAYS with the tube anode current very small.
I wasn't suggesting you do. I was merely pointing out that if you do monitor it with high resolution while slowly changing the voltage, battery DHT's sometimes show small jumps in current.
Operating tubes with constant current seems to be a bit of a fad in diyAudio. It is not wise.
It is not wise because the filament temperature is self regulating when the filament is fed from constant voltage. The same mechanism (Increasing temperature increases filament resistance. It's close to being directly proportional. With const voltage, increased resistance reduces power holding temperature down - there is thus a sort of negative feedback) that makes it regulating makes it tend to instability when the filament is fed from constant current. With const current and increase in temperature increases power, further increasing temperature - there is thus a positive feedback.
In another thread I pointed this out and a couple of people seemed determined to assert this is all nonsense, but it is basic physics. No profesional engineer would operate filaments or heaters const current unless forced to, e.g. in series fed undersea cable repeaters. But the tubes used are VERY special types.
Especially bad are folks who set their current regulators to the tube's rated nomimal filament current. The current that flows in or out of the filament pins in a DHT is of course affected by the emission current, so the tube filament gets overstressed right from the start. The current given in datasheets is by convention the current flows in teh cutoff condition.
You asked a question. I answered it.
You then stated that you are doing two things, a) underuning the filament, which may or may not be bad depending on the anode current, and b) operating the filament constant current, which is always bad.
I've given you good advice. It is entirely up to you whether or not to follow it. It's your tube. But if you don't want an answer, don't pose a question.
You then stated that you are doing two things, a) underuning the filament, which may or may not be bad depending on the anode current, and b) operating the filament constant current, which is always bad.
I've given you good advice. It is entirely up to you whether or not to follow it. It's your tube. But if you don't want an answer, don't pose a question.
I saw some advice in your first post . The rest was waffle . Undersea repeaters and all the other stuff you mention , OFF TOPIC and no relevance !!!
316a
I've been using starved filaments in my DHT preamps for many years with no issues. In particular I'm running the 4P1L preamp a tad hot for starved filaments at 550mA and anode +30mA. The preamp just had its 250 hours of operations (thanks to the counter) and haven't had any issues with the valves. Anyhow, 4P1L are cheap as chips. I'd only implement starved filaments for preamps, though. cheers Ale
Sent from my iPhone using Tapatalk
Sent from my iPhone using Tapatalk
Thoriated types need to follow the datasheet guidelines but I find anything with an oxide filament can often see improvements (especially WRT microphony) through starved or partially starved fils . Agreed re. 4P1L . I only back off a little for output stages and down to around 3.7V as drivers . I'm enjoying the 3B7 driver and feeling a bit more confident that I can tweak the filament volts up a little . 3B7 doesn't seem as noisy as some of the old B4 base 2V globes but it seems to be picking up some 100Hz which may be the filament supply . There is 0.9mV at the speaker which is quiet enough already and fairly decent for a breadboard all DHT SE amp
316a
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.