Finishing up Truepath system.

[ah693973]

I have been putting off finishing up my Truepaths for a few months now. I have several questions about the best way to finish them up and get them in their “final resting place”.
The system is proposed to look like this (going upstream):
1. Two JBL L5 speakers (6 Ohm).
2. Two Velleman 4700s for speaker protection.
3. Two Bridged Truepaths built with default values.
4. Twisted Pear Opus DAC. Balanced output based on Wolfson 8741 chip with usb or s/pdif input.
5. PC or media server source.
I guess the best way would be just to ask a couple of questions at a time and then see what the answer is.
The L5s are 6 Ohm and seem like they would push the Truepath pretty hard since I have 61 volts coming in from the power supply. I don’t really intend to drive them hard but went with the dual amps because it seemed cleaner than adding a balanced to single ended converter. If I replace the MOSFETs with IRFp4227PbF I should be OK right? I am aware that I will have to recalculate the current limiting and change from the default values.
I hate to have a relay (Velleman 4700) in the output path. Just seems like an easy failure point that could lead to degradation in the sound quality. I think that relying on the Truepath mute circuit will not protect your speakers from a failed MOSFET however. I believe the common failure mode for a MOSFET is ON and I don’t see how the mute circuit can help with that. One thing I am considering is having the relay from the 4700 short the output to ground and have a 10 amp fuses on the positive and negative supply lines from the power supply to the Truepath. I think this would work ok once you were up and running but the startup and shutdown on the 4700 have pauses to protect against speaker pop. Would the Truepath detect the short and not un-mute until the relay is open? Or would there be trouble (maybe something I haven’t thought of yet) with this scheme?

Andy

[thomaseliot]

I use K4700. These work perfectly and detect the 2,5V on output when the amp is muted and don't open. The problem is that at power off K4700 has no time to open the relays before a "swisch" of the power button.

The solution I adopted (hinted by Benn) is to use a double switch as power button and connect there either the SS power on and the mute switch. When turning off, the amp mutes, 2.5V appear on the output and the K4700 relays open before the power switch noise.

[busybee]

quote:Originally posted by ah693973


One thing I am considering is having the relay from the 4700 short the output to ground and have a 10 amp fuses on the positive and negative supply lines from the power supply to the Truepath.

Hi Andy,

As an alternative, I wonder if a solid state relay could be made to switch-in a low ohm shunt / dummy load across the outputs (without causing any trouble for the amp), this would be much faster than the relays on the K4700.

By the way, did you include the differential filter / common mode zobel network?

(The transducers I will be using have polyswitches installed, the idea is to set the over current protection at a threshold below the polyswitches' so to avoid the eventual loss of load. Also I try at length mentally to rule out the possibility of a MOSFET failure in the first place [8D] )

[ah693973]

Giovanni,
I will keep that mute switch “trick” in mind. It might be a prudent thing to do even if the turn off swish noise is cured some other way.

Busybee,
The shunt idea might work out well. I have some industrial Solid State Relay switches that I picked up for next to nothing at a surplus store that would do the trick. I would have to fuse or add a switch that would activate to shut down the power supply if the k4700 tripped. I ran a fuse between the filter capacitors and the Truepath on my first setup and things worked well. That might be the best solution.

I’ll show my ignorance here, but I’m not sure what you mean by the differential filter / common mode zobel network. Is that the output filter? If so, it is stock. Should I modify it?

-------------
I have ordered the IRFP4227PbF Mosfets and will be giving those a try. I am curious to see how much less heat is dissipated. One of the reasons I wanted to take on this project was to get rid of my class ab “heater”. All that wasted energy always bothered me.

Next question:
With a balanced connection between the Opus and Truepath, and the dual safety of a k4700 and the built in protection circuits of the 3020 chip, why are DC blocking capacitors necessary?

There seems to be a whole industry/religion built around the various input capacitors and their “sonic coloring”. I am sure there is some reason but I am unable to come up with it on my own.

Andy

[ah693973]

So I've got the Opus dac up and running and the two TPs operating in bridge mode but one of the TPs is muting if the other is started. One at a time things work great but if both are running then my second TP gets static and eventually goes mute. I jumpered the inputs together to see if it was interference in the input wiring but that didn't help. I guess the next thing will be to put each TP in a box to cut down on the static thrown off.

Has anyone run two (or more) TPs close to each other successfully? These are about 30cm apart on the bench and share a transformer and rectifier but have separate power supply caps.

It's always something.......

Andy

[thomaseliot]

I was waiting for your experiment

Let's think. What have the two Truepath in common? VPP, VNN, VN10, PGND, AGND. AGND is then in common with OPUS analogue ground. Undercurrent protection?

Volume control is on the PC?

Velleman K4700 relays open with only 1V DC (did you measure the DC at output?)

I'd like to watch to some pictures, including PS and wiring.

[jmanalo]

[quote]Originally posted by ah693973

Has anyone run two (or more) TPs close to each other successfully? These are about 30cm apart on the bench and share a transformer and rectifier but have separate power supply caps.
__________________________________________________ _______________

I haven't set all (4 boards) my TP yet, but running (one at the top of the other)two TPs and there is about 1/2 inch gap from the tallest cap to the second board. I haven't notice an anomaly so far. I've done the same thing on AMP5 and AMP10, both 8 channel (stuck 2 boards per side) in monoblock configuration but in single case. The only thing different was, I changed (per Jan's suggestions)the idle frequency caps on three boards.
On TPs, it's the same configuration except , I have not change the idle frequency cap, but have the caps in hand just in case.

Joe

[ah693973]

Giovanni,

I am trying to keep the setup as simple as possible to begin with. There is no speaker protection because I am using “throw aways” to test with. I am just using Windows Media Player (It controls volume) and the standard sound driver to drive the OPUS.

Have only had a chance to measure dc on one board TP1 (the one that works ok) and it needed a little tuning but was OK. Incidentally, I set each side to zero vs. the ground and then tuned it for zero between the two sides. I think that was the right way to do it?

I think you better be careful what you ask for with the photos:-) It is a big mess, everything is screwed down on two plywood pieces and there is no real thought, or order, to the wire runs. Here they are: http://picasaweb.google.com/andrew.m...idgedTruepath#
I’m not clear on the Undercurrent protection question. Could you elaborate?

Joe,
Thanks for the report, looks like maybe I have something wrong with my setup if you have things working with the TPs so close.

I think I will unhook the Opus and connect the two inputs on each TP to ground just to rule out trouble with the Opus. Other than that maybe make an enclosure for one or both TPs to kill the RF noise. Separate power supplies maybe? Build a little antenna and hook it up to my oscilloscope to see if the RF is significantly different between the two boards? It is always TP2 that makes noise and shuts down.

It is going to be difficult to spend much time on this as I have a job I am starting on Monday that is 2.5 hours away and will be gone from home 4 days a week and busy with family when I am home. So if nothing happens for a while it is not because of a lack of desire.

Thanks for all the help and advice.

Andy

[thomaseliot]

Hi Andy,

undercurrent protection in the chip happens when there is not enough current from the PS. Sourcing two amps with one transformer you have half the current avalaible for each one. Besides, in bridged mode, each amp draws twice the current, so the current avalaible for each rail of each amp is tranformer current/2/2/2 = 1/16. Your toroidal transformer looks like a 500VA 2X42V? If so it is not an issue at very moderate sound levels (500VA/60V/2/2/2=1A).

On the amp there are resistors for current sensing, maybe these are not the right values for bridged mode and your high voltage. There is a thread "Optimizing Truepath" where these settings are well described.

Thanks for the pictures. Opus-USB is not yet connected to Opus-Dac, so these are not the pictures of the working amp, but why grounds are not connected from Opus-Dac outputs to Truepath inputs?

[thomaseliot]

Any news?