Pmental,
Your PS choices sound okay. 21/24V, it does not make much difference. Clipping an amp is harder on a speaker, than driving it cleanly at more power, think twice before limiting dynamic range.
Your PS choices sound okay. 21/24V, it does not make much difference. Clipping an amp is harder on a speaker, than driving it cleanly at more power, think twice before limiting dynamic range.
I did not detect or even measure (on freq response) any evidence of heterodyning on my speakers with two Ybdz 2.0 amps running from a common supply. I am running them vertical (Left bass & treble on one amp and Right bass & treble on second amp). What would the herterodyning look like? A high HD peak at the beat frequency I would imagine? My only guess is that the modulation frequencies are so close between the two amps such that the beat freq is < 20 Hz (50 ppm for 400kHz mod rate - totally possible if quartz controlled osc is used), in which case I would not hear it as amp and speakers do not have much below 20 Hz.
The 24v at 1.5amps for full range and 24v at 4.5 amps for bass sounds like a good combo.
The 24v at 1.5amps for full range and 24v at 4.5 amps for bass sounds like a good combo.
Last edited:
Excellent, thanks again guys!
rsavas: I realize clipping is nasty, but I think bouncing off of Xmech on my woofers rather than clipping the amp is what is most likely to kill them and while I have no tools to measure my speaker cone movement I can limit amp power. Is clipping really that big of a problem on a woofers handling sub octaves in any case? I thought that was more likely to kill sensitive tweeters.
rsavas: I realize clipping is nasty, but I think bouncing off of Xmech on my woofers rather than clipping the amp is what is most likely to kill them and while I have no tools to measure my speaker cone movement I can limit amp power. Is clipping really that big of a problem on a woofers handling sub octaves in any case? I thought that was more likely to kill sensitive tweeters.
What would the hetrodyning look like?
Envelope (AM) modulation perhaps.
I guess we are talking two or more different cross-modulation sources. One is the PS/load noise getting cross coupled from one channel to the other(s) and then there is the actual test signal (audio BW) getting cross modulated. A few different test conditions to consider, cross-talk and PSRR.
Envelope (AM) modulation perhaps.
I guess we are talking two or more different cross-modulation sources. One is the PS/load noise getting cross coupled from one channel to the other(s) and then there is the actual test signal (audio BW) getting cross modulated. A few different test conditions to consider, cross-talk and PSRR.
Pmental,
I guess it a system issue, what happens first or is more damaging, speaker xmax limit or amp clipping?
Woofer xmax limit is not going to hurt your tweeter, but amp clipping will.
Yes clipping is known to be harder on tweeters than a woofer, well maybe not a woofer with a whizzer cone, but a separate smaller diam driver.
I saw in a Maxim design note, "REFERENCE DESIGN 4320" that they used a sliding HPF and some agc to limit power, thus the speaker xmax limit issue. It was made adjustable as well. This can be applied to a TPA31xx design as an added feature.
I guess it a system issue, what happens first or is more damaging, speaker xmax limit or amp clipping?
Woofer xmax limit is not going to hurt your tweeter, but amp clipping will.
Yes clipping is known to be harder on tweeters than a woofer, well maybe not a woofer with a whizzer cone, but a separate smaller diam driver.
I saw in a Maxim design note, "REFERENCE DESIGN 4320" that they used a sliding HPF and some agc to limit power, thus the speaker xmax limit issue. It was made adjustable as well. This can be applied to a TPA31xx design as an added feature.
September 2013 in this tread, what happened because sync did not work.
Is this hetrodyning? Chips working at a different switching frequency?
By definition heterodyne is the mixing of more than one frequency together producing sum and difference products due to circuit non-linearities?
So in above improper ckt op. of the slave not syncing up to the master was a form of distortion due to the hetrodyne effect.
So in above improper ckt op. of the slave not syncing up to the master was a form of distortion due to the hetrodyne effect.
Can't remember reading how tpa3116 chips "chose" their switching frequency within equal AM settings, 119or400khz seems also outside TI specs completely. But with same AM setting, all connected to gnd, which variables makes the specified 376khz to 424khz??? Table states eg T=25celsius, when using seperate amps for left and right I can imagine left ch and right ch temperatures could differ all the time, also load is specified above table R=4ohm, with anything other than mono the load will differ between left and right. Same for specified voltage, so ... ??? Syncing seems necessary ... ????
I am not syncing two boards with common supply and don't have any issues. 400khz is standard freq. You can select frequency depending on settings of the pins.
It's been a while since I last posted here, but I've been reading all along.
Guys... Finally it's done!
Linux based MPD - CS4389 "Lampucera" - Minidsp 2x4 - 2x TPA3116 + 3A Linear PSU + ~11.000uF + Scanspeak 18S4531G00 + D3004/6020 - DIY Signal / Audioquest FLX speaker wiring
The Ybdz's are singing beautifully!
It is the best system I've heard? No.
Is it the best scanspeak based system I've heard? No.
Is it better than some commercial scanspeak designs? Yes, it is. And some costing many, many times over. And paired together with so called hi-end electronics.
I am very satisfied with the TPA performance. Given the price, it's absolutely amazing. Most of the systems flaws I attribute to the DSP. I've heard the source and DAC with a Beyerdynamic DT 770, and I know some of it's qualities are missing. The extra ADC subdues some of the finer qualities of the sound.
Amplifier wise, the only thing I wish it could be better is major macrodynamic passages and transient response. They're are great, but when I remember a MBL Corona set paired with ELAC FS247 speakers I've heard I know (and wish) it could be better. But if you look at price tags, it's not an issue anymore. And perhaps a bigger transformer and capacitive bank could improve, it might not be the amp fault itself.
There is a lot of things to be done. Some more testing, some modding... But right now, all I want is to hear music!
Guys... Finally it's done!
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Linux based MPD - CS4389 "Lampucera" - Minidsp 2x4 - 2x TPA3116 + 3A Linear PSU + ~11.000uF + Scanspeak 18S4531G00 + D3004/6020 - DIY Signal / Audioquest FLX speaker wiring
The Ybdz's are singing beautifully!
It is the best system I've heard? No.
Is it the best scanspeak based system I've heard? No.
Is it better than some commercial scanspeak designs? Yes, it is. And some costing many, many times over. And paired together with so called hi-end electronics.
I am very satisfied with the TPA performance. Given the price, it's absolutely amazing. Most of the systems flaws I attribute to the DSP. I've heard the source and DAC with a Beyerdynamic DT 770, and I know some of it's qualities are missing. The extra ADC subdues some of the finer qualities of the sound.
Amplifier wise, the only thing I wish it could be better is major macrodynamic passages and transient response. They're are great, but when I remember a MBL Corona set paired with ELAC FS247 speakers I've heard I know (and wish) it could be better. But if you look at price tags, it's not an issue anymore. And perhaps a bigger transformer and capacitive bank could improve, it might not be the amp fault itself.
There is a lot of things to be done. Some more testing, some modding... But right now, all I want is to hear music!
Ramp3,
Great work there! I love the reuse of the VHS player case! We have similar setups, yes? I have different speakers but have to agree with you that it is one of the best if not best sounding systems I have heard. The price is laughable and that makes it all the better. You should glue the heatsinks on the amp - you may improve bass. Also, 3 amp linear supply is barely cutting it. Go for at least the 4.6 amp 19 volt laptop brick. I like how you flipped the miniDSP upside down to solder straight to the pins.
Here is my rig with same miniDSP and same amps.
Great work there! I love the reuse of the VHS player case! We have similar setups, yes? I have different speakers but have to agree with you that it is one of the best if not best sounding systems I have heard. The price is laughable and that makes it all the better. You should glue the heatsinks on the amp - you may improve bass. Also, 3 amp linear supply is barely cutting it. Go for at least the 4.6 amp 19 volt laptop brick. I like how you flipped the miniDSP upside down to solder straight to the pins.
Here is my rig with same miniDSP and same amps.
I have received my blue and black yung jing board today. it came with some supression x2 caps at inputs and also 1nF caps for bootstrap. I replaced with the right values but the board is dead.
Guess I got the trainee assembly.
Guess I got the trainee assembly.
Very nice! Your build looks much better. And actually it's a big old DVD!
I know 3A is not much, but it's what I had lying around. The 5A SMPS doesn't sound as good. When possible, I'll get two similar transformers, one for each. And I did leave the heatsinks above the chips for hours, they weren't even warm. So I didn't bother to fix them. But it wouldn't hurt to do.
I did want to get RCA plugs, but since good ones are costly. I had to resort to a McGyverism. Since you're a using a minidsp too, I am thinking about using a MiniDC. Good noise filter (used in car), and since it acts as a soft start, connecting to the amp mute function. And remove the redundant DC blocking caps at dsp's output as well.
I know 3A is not much, but it's what I had lying around. The 5A SMPS doesn't sound as good. When possible, I'll get two similar transformers, one for each. And I did leave the heatsinks above the chips for hours, they weren't even warm. So I didn't bother to fix them. But it wouldn't hurt to do.
I did want to get RCA plugs, but since good ones are costly. I had to resort to a McGyverism. Since you're a using a minidsp too, I am thinking about using a MiniDC. Good noise filter (used in car), and since it acts as a soft start, connecting to the amp mute function. And remove the redundant DC blocking caps at dsp's output as well.
These have been mentioned earlier in the thread... LM338K based adjustable power supply... I want to try these out as they're good for 5A [7A peak] through out their Vout range, just need the right VA rated transformer on the input side.
What is a miniDC? I just take some cheap old RCA to RCA cables and cut them in half and solder the flying leads to the amp input then plug the other end to the miniDSP. I found a source of cheap RCA PCB board jacks for like $0.50 a pair. If we want to take advantage of low noise capability of amp we should use preamps with balanced outputs.
I raised the question - what is the purpose of "syncing" the TPA3116 amps - on the TI support forum to the their engineers. Their answer went something like "syncing would control any beat noises generated which it can also be managed by modifying the layout of the traces on your PCB". First of all I do not know what they meant by beat noise. However, the second half of the answer suggested that the amps are located on the same PCB (may be sharing the same ground??) and are physically in close proximity. And then the person closed the response by saying that one can run all the amps (on the same board I presume) in master mode and it could be OK too.
In our case, the amps are on different boards and are quite far away from each other although they may be sharing a single PS. Anyway, based on whatever I can take away from the response posted by the TI engineer, syncing is not really that critical. I will be assembling a PBTL version using two "audiobah" boards (when they get here, soon I hope) running off a single 24 V SMPS. I will find out more. May be someone who is running more than one amp from one PS can comment on any anomalies detected?
Regards,