I feel compelled to repeat what I posted in #90, to be sure there is no potential for ground loops, you really should at minimum keep the high level and low level grounds totally separate. You might get away with combining the power and speaker grounds, but to use a single conductor for all grounds from the amp board to the power supply is inviting disaster. Even if you have a "star ground" on the amp board, you're still using one wire to connect it back to the power supply, combining low level and high level signals. I have the input, power and speaker grounds all on thier own conductors, never have any issues. Besides, like I said before, wire is cheap and it's easy to do it that way.
Mike
Power and speaker grounds should meet at the star ground for everything.
It is a bit confusing what you wrote about using one wire to connect back to the power supply. If you have the star ground on the amp board there is no need for more than one ground wire to the PSU, except that you're doubling resistance to have one ground wire handle return for both positive and negative rails.
One ground back to the PSU would not combine low and high level signals unless designing something odd where the low level audio signal goes through the power supply board traces but that is not the case in hurtz design nor most chipamps that I can recall, there would be a separate ground from the audio input, from the power supply, and from the speakers all meeting at a star ground on the amp board.
OK, where does the current from the speaker ground go, what's the return path back to the power supply? Now were's the ground return path for the low level side of things, the input ground? Looks to me like one common path, one single wire. I'll say it agian, you will get superior results, lower noise and distortion by keeping large current and small current paths back to the power supply separate. Even with the small value (10ohm?) resistor "isolating" the input ground, it is still using the same ground return path as the speaker. It might work your way, and I'm not trying to rain on anyones parade, but the way I'm suggesting works better and it's easy to do.
Mike
Mike
Another thought ocurred to me, when using one PSU for two AMPs, the "Grounds" are split on the PSU and join on each AMP seperately, could this pose a problem?
Also, the costs of the parts for one "set" (1x PSU 2xAMP) with 80VA 18V toroidal are around 60€.
cheers!
Glad to hear thatI'd be interested in a couple PCBs if everything goes well.
Also, the costs of the parts for one "set" (1x PSU 2xAMP) with 80VA 18V toroidal are around 60€.
cheers!
Download this:
PCB Via Current | PCB Trace Width | Differential Pair Calculator | PCB Impedance
it covers all the calculations such as current capacity you will ever require for a PCB.
thanks i was looking for this for a long time....
I'm glad you asked that, I took another look at it and realized you're using a single rectifier bridge, whereas I normally use two with a dual secondary transformer to maintain independant supplies for each polarity, I should have been more observant. So it would be better to use the original design where the ground trace was a continuos ring around the outside of the board. However, for that design I would personally prefer to have a single large ground trace down the center of the board and put the power rails on the out side, the way you have it is sort of "inside out" to my eyes. Sorry for the misdirection and confusion.
Mike
Mike
So, the first board is finished, but to my disappointment I must say the sound is extremely muffled/dull in comparison to my Symasym. I somehow suspect Cin might be the cause? I used 6.8uF instead of 4.7uF... enzoR said he didn't encounter any problems and he left the following components unpopulated: Cin, Cn, RnL, Rb2, Rz, Cz...
cheers!
cheers!
Images of the complete build would be helpful. What speakers were used to generate subjective results between the discrete amp and could you use the same power supply? Did you fiddle with the grounding choices discussed? Cin's value is unlikely the least of yer worries. PS I thought qualitative testing results were promised.
On a second note, I think the resolution of the sound is quite well, it just sounds as if a bad equalizer has been set to maximum on bass gain. (it's not the source).
I would also suspect that Cn with 100nF and Cd with 100nF and then again Cc with 220pF is a bit overkill on RF-filtering. However the amp is absolutely silent.. no noise/hum whatsoever on the speakers.
I would also suspect that Cn with 100nF and Cd with 100nF and then again Cc with 220pF is a bit overkill on RF-filtering. However the amp is absolutely silent.. no noise/hum whatsoever on the speakers.
test gain flatness over 20-20K, I generally find it fruitless to discuss technical circuit design choices based purely on one persons esp the PCB disigners subjective impression. IMO the power supply is responsible for most "quality" results in the bass power band, but if it sounds bass heavy yours might be too good. lol
Last edited:
Heres the image, the test setup is mono with short leads from amp to psu to toroidal. I (of course) used the same speakers for testing. the same PSU can't be used since the symasym is with an integratedd psu. The grounding is the same as on the Layouts I published, no changes have been made since then.
I don't have the testing results myslef yet, enzoR only gave me a short notice that everything is fine, and that he's off to vacation.
since you mention the same PSU, I'm not talking about a slight dullness, this sounds really bad, even without comparison to anything else.
cheers!
I don't have the testing results myslef yet, enzoR only gave me a short notice that everything is fine, and that he's off to vacation.
since you mention the same PSU, I'm not talking about a slight dullness, this sounds really bad, even without comparison to anything else.
An externally hosted image should be here but it was not working when we last tested it.
cheers!
test gain flatness over 20-20K
How would I do that?
cheers!
P.s. too good a PSU? D'Oh! althoug 14.700uF per rail(7.350uF per channel) doesn't seem that much compared with some other PSU's I've seen. However 10.000uF per channel seems more common.
You could use a function generator and decent AC meter, or O-scope for old school or even some loopback test on a everyday PC sound card. generally just sweep the frequency over till you find the upper and lower -3dB -0.707 Vo from center, and there shouldn't be any significant change between those 2 points ie flatness/.
Last edited:
Test setup:
no function generator... no o-scope... will research loopback thingie..
An externally hosted image should be here but it was not working when we last tested it.
no function generator... no o-scope... will research loopback thingie..
make sure you use an accurate 30dB or so attenuator from the test load to yer sound card. dont want to blow up yer PC or anything. First test the attenuator for the baseline. watch / calculate the levels
I suppose you could use your speaker in a pinch for the test load at much reduced levels cause it may stress the woofers at lower frequencies and the ears up higher.
I suppose you could use your speaker in a pinch for the test load at much reduced levels cause it may stress the woofers at lower frequencies and the ears up higher.
Last edited:
Ok quick test with RMAA gave me (10k/1k resistor,otherwise input level was too low):
THD: 0.021%
and freq response:
THD: 0.021%
and freq response:
An externally hosted image should be here but it was not working when we last tested it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Chip Amps
- The yet tiniest single-sided LM3886?