need suggestions for an AB amp project

[meanman1964]

Originally Posted by boywonder

How much 40-0-40 VAC current do I need for the Honey Badger? I see this board in the store so that's appealing.....

I like the idea of using the 25-0-25 VAC output for another lower power stereo output......I'll have to research MyRef-FE and TDA7293.

Can you provide more details on the RC snubbers?

I usually hang out in either the tubes forum or the Pass Labs forum so these projects are all rather new to me.

Thanks!

With the 2x25VAC you can build the FetZilla

[boywonder]

Everyone: thanks for all of the input!

Lazycat: I checked out the VSSA thread; sounds interesting......I hoping to find a project with PCBs available so I will stay tuned.

Next (dumb) question. Can I estimate the current capability of the 40-0-40 taps with any accuracy by knowing that the AV receiver was rated at 100W output on the main stereo channels? I understand that there is a spectrum of possible class A and AB transition points based on idle current. I would also assume that this operated in vary shallow class A since it was an AV receiver.

Also, if I assume that the 25-0-25 taps were powering the center and rear channels (35W each) can I get a meaningful estimation of current capability for these taps?

Should I just test the transformer under load? I have a few large power resistors but in very limited values.

[AndrewT]

Can you see and therefore estimate the wire thickness used for each of the 18Vac, 25Vac and 40Vac taps?

Winding wire thickness is a VERY good guide to current rating.
Assume 3.1A/sqmm of cross-section.

[boywonder]

Originally Posted by AndrewT

Can you see and therefore estimate the wire thickness used for each of the 18Vac, 25Vac and 40Vac taps?

Winding wire thickness is a VERY good guide to current rating.
Assume 3.1A/sqmm of cross-section.

All of the wires are Hitachi brand 22awg 105 deg C marked.

[ilimzn]

Originally Posted by AndrewT

Can you see and therefore estimate the wire thickness used for each of the 18Vac, 25Vac and 40Vac taps?

Winding wire thickness is a VERY good guide to current rating.
Assume 3.1A/sqmm of cross-section.

Actually, by measuring the unloaded voltage and the resistance of the winding (short at least one other winding as some multimeters will have a big problem with the inductance of the winding), can give a good estimate of the available current.
Standard EI transformers are wound at up to 10% regulation, which basically means the voltage drop from unloaded to fully loaded is 10%. Using the measured voltage and resitance it's a matter od using ohms law to estimate the available current, and this would be step 1.
Step 2 is adding together voltage x current of all windings to get an estimate of the VA rating of the transformer. This is then compared to the core size. A quick estimate for standard EI cores would be to calculate the cross-section of the central part around which the bobbin is mounted, in square cm. Take 90% of this number and square it, and you get a rough estimate of the VA rating. Divide that with estimate you got from the voltage and resistance measurements, usually this latter will be higher because a rather pessimistic figure was used for regulation, and you will get a figure less than 1, 0.7-0.8 would be quite common. Multiply your estimated winding currents by thet number and you will likely be very close to the actual spec, with some safety margin.

So, here is an example.
Suppose you measure a winding at 40V unloaded. The winding resistance is measured at 1 ohm (remember to subtract the resistance of probe wires which you get by shorting the probes - not the definitive method but a sufficiently precise ballpark value).
Assuming regulation is 10%, this resistance will drop 10% of 40V, i.e. 4V at full load, ohms law tells us this will require 4A of current. So, this winding is capable of sourcing 40V x 4A = 160VA of power. For multiple windings do the calculation for each and add the VA contributions together.
However, measiring the cross section of the iron core inside the bobin gives us say 12 square cm, but since not all of it is iron - there is some insulation and air, assume 90% is iron, and so the actual cross section is 10.8 square cm. THis in turn when squared gives us 10.8 x 10.8 = 116.64VA as the VA rating of the core. Now we divide this with the forst estimate, 160VA and we get 0.729. We use this to multiply our first current estimate to get something closer to the ctual current, in this case it is 4 x 0.729 = 2.916A

This estimate will usually fail for very high current or very low voltage (difficult to measure low resistance of the winding) and very low current (winders will often use thicker wire to make winding easy if this is not the largest power winding on the core) windings in that it may not represent how the transformer was originally designed, but you could still source the calculated current.

[danielwritesbac]

Originally Posted by boywonder

Say 25-60W or so? We don't need much horsepower.

(60W*2 for lower ripple)*2 channels=240va. Okay! Right ballpark!
Yes, it is possible to get the 60W per channel stereo, in high fidelity. It will have an additional higher capacity of midfi useful for headroom management that sounds better than clipping. If the power supply capacitance is sized big enough to outlast a bass beat yet small&fast enough to recover before the next bass beat, it may do 75 watts per channel if the transformer allows, and also possibly rather large bass peaks that can vary in quality. Surely using the Honey Badger makes a better quality amplifier; however, the Denon's smaller power range behavior persists because of using a lower amperage, mass market retail transformer. This behavior can be quite useful if you have non-gigantic 8 ohm speakers.

Result:
Honey Badger boards with 2x MJL parallel for outputs.
Do C3,C4 as 220u||220u and right size the input cap.
5x6800u per rail simple split rail supply with KBPC3502.
Due to retail transformer, avoid using 4 ohm speakers.

[AndrewT]

Originally Posted by boywonder

All of the wires are Hitachi brand 22awg 105 deg C marked.

That reads like the leadout wires.
It's the wire in the winding you need the thickness of.

[boywonder]

Thanks for the replies everyone!

AndrewT: Yes those are the lead wires, I'm not sure I want to dig into the case...

ilimzn: I'm familiar with the concept you've outlined, and used a handy chart published years ago to estimate the current capability of an unknown tube power transformer based on winding resistance. In this case, even with my Fluke 79 meter, it's difficult to measure 1 ohm or less even on the sensitive R range. Tube power transformers at least have resistance that can be measured with a typical meter.

Daniel: So one can scale the number of outputs on the honey badger without much other changes? ie front end stays the same, etc.? This sounds do-able....

[danielwritesbac]

Originally Posted by boywonder

Daniel: So one can scale the number of outputs on the honey badger without much other changes? ie front end stays the same, etc.? This sounds do-able....

Yes, very scalable. The Honey Badger is the result of a modular amplifier system that Ostripper has been designing for some time. It is an advanced version of the AX, part of the Mongrel amp (modular options) series. The MONGREL (supersym II)

[boywonder]

Thanks Daniel:

So, what outputs (and how many)?

I'm assuming that running less than 3 pairs at higher bias currents makes more sense than fully populating the outputs and running them with little bias due to transformer current limitations.

I have been searching the various Honey Badger threads about output devices and bias current, but have only found spotty info. I see that most folks are using 600-800VA transformers for this project, and I'm sure the one I have falls far short of that (comparing it to the size of the 400VA Antek 20V transformer in my F4).

I'd be plenty happy with anything between 25-75W or so of output, and happily sacrifice max power out for hotter bias with less/smaller outputs. This project is sort of a "scrap box challenge".