Basic "safety" question re P-P test amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm a tube guy really, but understand the principles of transistors. I want to build a small test amp of known low(ish) distortion to get some measurement of speaker behavior to compare with the higher distortion of the tube amp I'm building.
I know how to build the general sort of amp in this link: Class B Audio Amplifier | Analog Integrated Circuits | Electronics Textbook.
I've built them in the past as instrument (as in "scientific") amplifiers. But I'm nervous about the direct DC connection of the speaker. Is this safe? I don't want to risk burning up my speakers.
Thanks...
 
In general: that DC connection is a risk, which is why most amplifiers have some sort of DC protection.

About the schematic you referred to: when the wiper looses contact, there is no way of predicting what will happen to the speaker. On top of that, there is no bias current so you'll get crossover distortion.
 
Based on the text:

" LEARNING OBJECTIVES

How to build a “push-pull” class B amplifier using complementary bipolar transistors
The effects of “crossover distortion” in a push-pull amplifier circuit
Using negative feedback via an op-amp to correct circuit nonlinearities"

I would not use this as a "small test amp of known low(ish) distortion".

Have you been here yet?:

DIY Audio Articles
 
Founder of XSA-Labs
Joined 2012
Paid Member
You can always cap couple your PP solid state amp with suitably sized large back to back electrolytics. Say, two 6800uF 63v will give equivalent 3400uF that will give enough bass extension.

A good easy to build PP amplifier of known low distortion is probably a chip amp like LM3886, etc.

DUG’s suggestion to make something from ESP is spot-on.
 
For test and comparison, do you need a high power amplifier?

I suggest a small push pull ClassA biased single pair amplifier running on +-20Vdc supplies and biased to give max output of 2Apk in full ClassA. That's just 40W of dissipation to the heatsink and a maximum power into 8ohms of ~ 16W
You could even bias to 600mA, or 700mA for a smaller heatsink. 600mA bias gives ~ 1.2Apk in ClassA and max power ~ 6W into 8ohms.

When built, thoroughly test with and without any dummy load. Then do lots of cold starts into a cheap speaker load. Repeatedly check DC output offset, cold, warm & hot.
Once you are completely sure the new build is behaving, recheck again and then connect your expensive speaker.
 
Last edited:
You can rearrange the power supplies of the amp in post 1 to protect the speaker with one 3300 uf snap in cap. Make the negative supply the speaker ground. Make the center "ground" a +12 and the "ground" on the op amps. Make the +v a +24. The input already has a cap to shift the voltage. The output cap protects your speaker. Cheaper radial lead 1000 uf passes a little less bass.
In test, back to back electrolytic caps like post 4 sound bad at 1.5 W even with a great amp driving it. The caps do something chemical at 0 volts crossing over. Audible mainly on high frequency signal like bells or top octave piano.
This circuit will sound funny. Build a 6 transistor AX6 to improve sound a lot. Already has a speaker cap and single supply. www.diyaudio.com/forums/solid-state/236256-retro-amp-50W-single-supply-20
Or a TGM8 is very similar, bigun says you can get the boards from a custom house for $25 or so. www.diyaudio.com/forums/solid-state/314330-class-ab-amp-recommendations
There is also a thread on chip amp discussing how to drive pass transistors with an op amp without crossover distortion. I don't have the url handy. It has a few more resistors and capacitors than the example post 1. An op amp with more drive current like RC4560 will produce slightly more power.
Have fun experimenting
 
Last edited:
Thanks to all. Main confirmation is that DC connecting this kind of thing to my good speakers isn't a good idea out of the gate.

I know how to make these amps because I used them as servo motor drivers. As a result I've got a pile of TL082 op-amps and PNP/NPN complementary pairs (or a few sorts, I forget). Also a stout +/- 20V power supply.

Closing the feedback loop around the transistors in class B mode gives "good enough" crossover to not disturb a servo drive. The TL082 anyway has a pretty fast slew, but I could see a little "dogleg" around zero V as I recall.

Would sticking with (roughly) the design I'm thinking of but biasing it into A/B be enough for audio testing, aside from my other concerns? Full class A would probably require bigger heatsinks than I have in my junk box.

Moving the supply all positive would work too, and I have some 10,000 uF 63V caps I could use for coupling. Again I'm not looking for something spectacular, 10W would be more than enough. Mostly a pretty good test amp to test speaker impedance, tube output transformer characteristics, that sort of thing.

Thanks again. I think I've learned enough to go ahead and try some stuff anyway, after all it's a hobby not a business...
 
Also see post 30 of this thread: BD139 / BD140 ? for an op amp driven amp. No reason TIP41/42 couldn't sub for the BC139-140, but Bruno lived in Slovenia or someplace and couldn't get them. The drivers are just 500 ma TO92 transistors, pn2222/2907 might work, 2n5401/5551 surely would, and I like MPSA06/56 which are also good for higher voltage work and cheaper than 2N5401/5551.
 
I know how to make these amps because I used them as servo motor drivers.

Closing the feedback loop around the transistors in class B mode gives "good enough" crossover to not disturb a servo drive.
The "dogleg" around zero V .
The dog leg around zero.....
Here is the key point. You will never get rid of it with feedback.
That is what makes power audio amps a very different story than servo motor amps.
 

PRR

Member
Joined 2003
Paid Member
....I'm nervous about the direct DC connection of the speaker. Is this safe? I don't want to risk burning up my speakers....

Everybody does it.

In a stereo amp, it means 2 large caps instead of 3. Also a little less power-robbing and bass limit.

When everything is working right, there is "NO" DC on the speaker. (Typically a few milliVolts, "nothing" to a speaker.)

Many-many low-price amps do not add a DC monitor/shutdown circuit. In fact that was fairly rare (or problematic) until recently.

Yes, IF you melt an output transistor you get near full rail DC on the speaker, which is typically more than it can stand.

It is not a "bad" idea. If you are paranoid, you can find many other possible faults from tweeter-frying ultrasonic squeal, woofer-slamming motorboating, to transformer breakdown and fire.

I've built and fixed many direct-coupled power amps and have not yet burnt a speaker.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.