Lowering B+ to 68V and working from a 34K primary impendence get's us closer I think. This is with a 70% idle bias. Still creeping across max dissipation, and getting close to the correct primary impedance from a Hammond 125B is now more confusing.
Just a suggestion, work with the primary impedance you have available, and then select your operating voltage and bias accordingly. Use 22.5K Primary for 8 Ohm Speaker wiring as a reference point for example. That is the max impedance they show on the table for that transfo with an 8 Ohm Speaker. Also per the 5672 data sheet the typical load resistance would be 20 K ohms, so it would seem OK. From there you determine the load line. I would normally then consider the transfo max current and voltage, but in this particular case you will be way under it's rating for wattage, and your DC bias limit will be way under the transfo rating.
Here is another suggestion: Setting aside overdrive for now, just play through the existing circuits and turn down the volume until it is at a level you would find acceptable. Now measure the output. That becomes your power target for the eventual solution. RAther than hoping/assuming 0.1 watt is the answer, find out.
I'm sorry. I thought this was an audio amp to be listened to. You want to watch the loadlines? I've done that, amused myself, nobody else.
Must agree with shanx: "....work with the primary impedance you have available, and then select your operating voltage...."
At this part-watt level, dropping plate supply voltage is not a major power station but a fistful of cheap 500r 10W resistors.
I like the project. Very curious how it will sound. About the Hammond transformer: you could wire it to get 44k primary impedance with 8 ohm load. Then with let's say 80V anode voltage the tubes would be ok it seems. Not much power, about 150mW, but that's what you want, right?
I've been thinking about this amplifier idea for a while and going round in circles. I'm considering the power supply - are there any obvious elegant solutions to supply the quite different voltage requirements of three 12AX7 and two 5672?
As I understand it, the 'traditional' approach for these type of amplifiers involves a transformer with a tap to supply heater voltages. The 5672's take 1.25V, and the 12AX7's would typically be supplied with 6.3V, so this is the first consideration.
Next, ~300V DC B+ is sent to the pre-amp section, whilst perhaps ~70V DC B+ would be required for the power amp section.
As I understand it, the 'traditional' approach for these type of amplifiers involves a transformer with a tap to supply heater voltages. The 5672's take 1.25V, and the 12AX7's would typically be supplied with 6.3V, so this is the first consideration.
Next, ~300V DC B+ is sent to the pre-amp section, whilst perhaps ~70V DC B+ would be required for the power amp section.
For pentodes, the load line is unnecessary. The optimum load resistance is close to the DC resistance Vak/Ia. Several books of the tube era confirm the same. Enjoy low power devices.
Some time ago I built this:
[/youtube]
[/youtube]
It uses russian triodes in the preamp and PI and two 5672 in PP. It was based on the liverpool amp. I guess you could make it more bassman like.
Even with a non-guitar speaker it is loud. Not as loud as a 1W amp, but loud enough for the person in the next room to complain.
I can't remember the voltage I used...have to check if I posted it somewhere else.
[/youtube]
It uses russian triodes in the preamp and PI and two 5672 in PP. It was based on the liverpool amp. I guess you could make it more bassman like.
Even with a non-guitar speaker it is loud. Not as loud as a 1W amp, but loud enough for the person in the next room to complain.
I can't remember the voltage I used...have to check if I posted it somewhere else.
So, just to add,
The output tubes are biased at -7.7v, with an elevated heater, using 44R and 62R resistors.
It works nice with the 22.5k Hammond transformer. Voltage at the Transformer is 87V.
The output tubes are biased at -7.7v, with an elevated heater, using 44R and 62R resistors.
It works nice with the 22.5k Hammond transformer. Voltage at the Transformer is 87V.
Hi Thomas, thank you - that looks pretty similar to what I'm hoping to build.
Can I please ask about the voltages if you can remember? You are using 6N17B and 6N16B in the preamp/PI, are you sending 6.3v to the heaters for these? And then 87V to the plate/anode for both the 6N17B, 6N16B and the output 5672s?
Can I please ask about the voltages if you can remember? You are using 6N17B and 6N16B in the preamp/PI, are you sending 6.3v to the heaters for these? And then 87V to the plate/anode for both the 6N17B, 6N16B and the output 5672s?
Heaters of the preamp tubes are in series and powered directly from the 12V DC supply.
Max. Voltage is 87V at the Transformers' centre tap. I have not measured the voltages directly at the anodes, but they drop according to the resistors used. Preamp tubes are much lower voltages as a result.
To achieve that I changed the SMPS resistors (560k and 10k), originally they are 220k and 1k, but that gives around 250V.
The heaters of the 5672 are powered in parallel, with resistors in series to drop the voltage from 12V. Instead of a single resistor, I used two, so that the voltage at the negative side of the filament is roughly 7.7V.
Max. Voltage is 87V at the Transformers' centre tap. I have not measured the voltages directly at the anodes, but they drop according to the resistors used. Preamp tubes are much lower voltages as a result.
To achieve that I changed the SMPS resistors (560k and 10k), originally they are 220k and 1k, but that gives around 250V.
The heaters of the 5672 are powered in parallel, with resistors in series to drop the voltage from 12V. Instead of a single resistor, I used two, so that the voltage at the negative side of the filament is roughly 7.7V.
In case it's helpful, here's the one I built along similar lines, using 5672 output tubes (I also learned from 5672 builds Thomas did):
I'm running the output tube bias a little hotter than Thomas, with 6.5V on the + side of the filaments and 5.25V on the - side (1.25V across). I wasn't clear on whether the -6.5V grid bias on the datasheets is meant to be in reference to the + side, - side, or average voltage of the filament/cathode, but what I have sounds good to me.
If you want to use 12AX7s and you have a 6.3V tap for their heaters, you could probably use that for the 5672 filaments as well, with an ~50R resistor to ground from the - side to get the filament voltage right (filaments are roughly like current sources: 2 parallel 50mA filaments means 100mA * 50R = 5V on the low side of the filaments). Those 6.3V taps are often really 6.5V or more anyway, so the bias may be closer to right (but you may need a few more ohms of resistance).
Getting the 70V for the 5672 plates from a transformer supplying ~300V for 12AX7s is probably just a matter of figuring out the right value for a voltage-dropping resistor and a reservoir capacitor.
I'm running the output tube bias a little hotter than Thomas, with 6.5V on the + side of the filaments and 5.25V on the - side (1.25V across). I wasn't clear on whether the -6.5V grid bias on the datasheets is meant to be in reference to the + side, - side, or average voltage of the filament/cathode, but what I have sounds good to me.
If you want to use 12AX7s and you have a 6.3V tap for their heaters, you could probably use that for the 5672 filaments as well, with an ~50R resistor to ground from the - side to get the filament voltage right (filaments are roughly like current sources: 2 parallel 50mA filaments means 100mA * 50R = 5V on the low side of the filaments). Those 6.3V taps are often really 6.5V or more anyway, so the bias may be closer to right (but you may need a few more ohms of resistance).
Getting the 70V for the 5672 plates from a transformer supplying ~300V for 12AX7s is probably just a matter of figuring out the right value for a voltage-dropping resistor and a reservoir capacitor.
Last edited:
I stupidly didn't consider that your 6.3V tap will be AC. You'd need to rectify that to DC for the 5672 filaments (AC would be very noisy on a direct-heated cathode), and after rectification you'd have 7.5VDC or more to work with. Resistors before and after the filaments can then put their bias wherever you like.
Sounds sensible, thanks for the heads up.
I was wondering whether it might be useful to look for a power transformer with a suitable secondary bias tap that could be rectified and used as B+ voltage for the 5672 output stage? For example, the Hammond 369BX.
With a 369BX with the centre tap earthed, I was thinking that rectifying the 50VAC bias secondary would get ~70DC. Would this mean though that this would reduce the available high voltage secondary to 150V per side, so a beefier transformer than this might be required if I want to get voltages of ~300VDC for the pre amp tubes?
I was wondering whether it might be useful to look for a power transformer with a suitable secondary bias tap that could be rectified and used as B+ voltage for the 5672 output stage? For example, the Hammond 369BX.
With a 369BX with the centre tap earthed, I was thinking that rectifying the 50VAC bias secondary would get ~70DC. Would this mean though that this would reduce the available high voltage secondary to 150V per side, so a beefier transformer than this might be required if I want to get voltages of ~300VDC for the pre amp tubes?
Last edited:
In reference to my transformer questions, I was thinking that a 370CX might be closer to my requirements:
Hammond 370CX Power Transformer
275-0-275 @ 75mA
59VA, 550V @ 81mA, 50V BIAS,
Filament 1- 5V C.T.@ 0.6A, Filament 2 - 6.3V C.T. @ 2.5A
Hammond 370CX Power Transformer
275-0-275 @ 75mA
59VA, 550V @ 81mA, 50V BIAS,
Filament 1- 5V C.T.@ 0.6A, Filament 2 - 6.3V C.T. @ 2.5A
Yeah, I think you could get 70VDC from a 50VAC bias tap with a half-wave rectifier and a big filter capacitor. (Full-wave rectification is tricky with only the one bias tap if you also need the CT grounded for the 300V rectifier. Though it might work okay to do a bridge rectifier for the 70V and have the 300V rectifier use the bridge's ground instead of the CT. I haven't thought about it enough.)
That said, I wouldn't spend too much more on a transformer with a bias tap when (eyeballing typical plate + screen current from the datasheet...) an ~56k 5W resistor using an extra watt or so of power would probably do the trick. I guess one worry in that setup would be the 5672 plates seeing 300V in the tiny fraction of a second while their filaments warm up. Those filaments warm up pretty much instantly, but I'm not 100% sure whether that's likely to be a problem.
That said, I wouldn't spend too much more on a transformer with a bias tap when (eyeballing typical plate + screen current from the datasheet...) an ~56k 5W resistor using an extra watt or so of power would probably do the trick. I guess one worry in that setup would be the 5672 plates seeing 300V in the tiny fraction of a second while their filaments warm up. Those filaments warm up pretty much instantly, but I'm not 100% sure whether that's likely to be a problem.
Last edited: