12au7/ fujtsu mb3732 headphone amp

[preludefan]

hi all,
im quite new to the amp building scene on the net, i have a few project on the go for vairous, mainly tube amps, however this one is a little different and i have a few questions for you all if you would be so kind to awncer.

i am planning to go working overseas sooner or later next year and all i will be taking is my clothes, tools and laptop, i love music and cant bare to leave my tube amps behind for months at a time, so i thought id make a small improvisation, the plan is to build a hybrid tube preamp/chip power amp headphone amplifier.

i am going to use a 12au7 tube as these have been known to work ok at around 12v- i would go to a 12ax7, but my budget is low for this project and i wish to use a toridal transformer with a 15v supply that is available to me cheap, to have a custom made one to give the needed 180v to a 12ax7 would blow the projects budget.


the amplifier is going to be fed from a toridal transformer powered supply.

i have an old car stereo removed from one of my project hondas that uses 4 mb3732 power amp chips and they put out around 14w each per channel into 4 ohms, now this may sound overkill but the distortion levels at the volumes id be listening to this through headphones on would be almost nothing, and correct a tube head if hes wrong, but with electronic amps, the larger your amp to a point, the lower your THD right? so i thought using two of those might work out ok?

i was lucky enough to find the datasheet for this amp online:
MB3732 pdf, MB3732 description, MB3732 datasheets, MB3732 view ::: ALLDATASHEET :::

could a few of you take a look at this and tell me what you think and if it is worth using this chip, and how it would stack up against a mosfet amp instead?

also, is there any suggestions any of you have for the power supply, im going to be using a torid transformer as mentioned, but is it best to use a bridge rectifier with caps alone or am i better to regulate it from 24v down to 15v?

thanks in advance

mike

 

[sgrossklass]

BTL amps aren't going to be of much use since headphones typically have their driver ground connections tied together.

Even if they weren't BTL, using power amps would still be tricky. One would have to come up with a nested feedback setup to keep noise and distortion at bay (these things are hissy). Nested feedback isn't exactly for beginners.

If you're stuck with relatively low supplies, I don't really see the point of using a tube input stage. Anyway, I'd go for a cathode follower with a healthy amount of anode current here, that would be nice for driving a relatively low-resistance volume pot (anywhere from 2k2 to 10k).
Then you could use whatever kind of low voltage noise headphone amp/driver circuitry you like.

Do you intend to go split supply or stick with single? Is the transformer a single secondary 15V~ job, or does the secondary have a center tap?

Even if it's a single secondary xfmr, I'd try and go split supply, à la how the O2 does it. The NJM2068 opamp used there wouldn't be a bad idea in terms of voltage gain either (as would the gain setting resistor values). You could then beef that one up with a nice discrete AB output buffer within the feedback loop, à la Eaton amp (BD139/140 class driver transistors at ~5..10 mA bias would be useful, BC337/327 may do if the xfmr is somewhat wimpy; dunno 'bout the MOSFET side). Output series resistor maybe 1 ohm.
Distortion performance should be fine at moderate gains like 3x..5x.

 

[preludefan]

thankyou for that info, very interesting indeed.

since the first post i have done extensive testing with this amp and it does not appear too hissy- white noise is there, but it is very low.
also i tested the amp with headphones and with each speaker sharing the same neutral- it does not appear to affect anything at all, so long as the negative from the headphones is isolated from the negative of the supply.

the toroid is a 15v 0v 15v. so i will go down the split supply route.

mike

BTL amps aren't going to be of much use since headphones typically have their driver ground connections tied together.

Even if they weren't BTL, using power amps would still be tricky. One would have to come up with a nested feedback setup to keep noise and distortion at bay (these things are hissy). Nested feedback isn't exactly for beginners.

If you're stuck with relatively low supplies, I don't really see the point of using a tube input stage. Anyway, I'd go for a cathode follower with a healthy amount of anode current here, that would be nice for driving a relatively low-resistance volume pot (anywhere from 2k2 to 10k).
Then you could use whatever kind of low voltage noise headphone amp/driver circuitry you like.

Do you intend to go split supply or stick with single? Is the transformer a single secondary 15V~ job, or does the secondary have a center tap?

Even if it's a single secondary xfmr, I'd try and go split supply, à la how the O2 does it. The NJM2068 opamp used there wouldn't be a bad idea in terms of voltage gain either (as would the gain setting resistor values). You could then beef that one up with a nice discrete AB output buffer within the feedback loop, à la Eaton amp (BD139/140 class driver transistors at ~5..10 mA bias would be useful, BC337/327 may do if the xfmr is somewhat wimpy; dunno 'bout the MOSFET side). Output series resistor maybe 1 ohm.
Distortion performance should be fine at moderate gains like 3x..5x.