Diy-dac 1543

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Now for some feedback

I have got the dac up and running, with the EC mod, removed the input diode and dumped the power LED. Replaced the cap on the dac chip with 1000uf and a small smd bypass.
It is feed by a 7.2AH SLA.

I tried it without the output caps, resistors after the caps and the 2200p caps and was getting distortion on the mid freq peaks, so I put them back with new 3.3uf polyseter caps, however it now appears the problem was over driving the tweeter amp modules (they were running on 12v supplies). When I changed the tweeters back to 24v supplies the problem was gone.

I am still getting some grain so looking for some answers there, but I only have 1 1/2 hrs up on the dac so far so maybe that is an issue

I am now looking to refine the power supply of things, separate raw dc supplies for all chips, so 5v for the dac, 3.6 for DIR9000 and 5v for the TL chip. I am assuming I will need to use a set of relays to synchronise the turn on.

My take at present, very detailed mids, big bold dynamics and strong bass, highs are fine, quite analogue but a little too much grain at present.

Ideally strange as it might seem the output is a little too high for my system, I have the L Pads cranked right back just to keep the listening level room friendly (oh the problems one has when your speakers are 100db plus, actually the tweeter hits 108). This I guess can be solved with more resistance on the speaker lines, I am loath to throttle the signal from the dac, I have tried running through the active pre but basically it kills the sound stone dead in comparison.

The other option is change the gain on the modules, overall though things are looking positive, just fine tuning.
 
Folks, I did it!

After my wife drove away with a friend and the children felt asleep, I went to my tinker room in order to mod that red HK PCB.
The solder iron got 350°C and I took a look into my tinker box.
I removed the green WIMA (fake?) 2200pF caps, the 10µF output caps and replaced it with 10µF Intertechnik Q4 caps. And I paralleled the I/V resistors (1k) with 7.87k so I got 886 Ohm for them. Ah yes and that green decoupling cap at the TDA1543 changed for a Panasonic FC 100µF.

I didn't expect very much but than I was really surprised! The sound was much better than before! There were dynamics and highs (not the sparkling ones but nice). I was quiet satisfied with the mods. The only downside is the narrow sundstage. WIth my other gear the stage is 1-2 meter wider. But perhaps it will become better.


@Zero one
Seems a little bit strange that output level is too high. EC reported that with his mod (3V, no pin 7) the output level is just 1.5Vpp and that isn't much at all. With that output you have to crank your vol pot to 3 o'clock!
 

Attachments

  • dacmod.jpg
    dacmod.jpg
    96.1 KB · Views: 617
Hi Tolu

My application is a bit outside the norm, the DAC drives the amp modules directly and the volume is controlled via L pads on the output side. The amp modules will clip at around 600mv on +/- 24v, but the tweeter modules can run on 12V, they will clip at a far lower level on this input voltage, but they sound sweeter too for this specific application run on 12V supplies. Output power is no issue even on 12 volts as the drivers run from around 104--108 db per watt.

But to keep the amp module from clipping on 12 volts with the input from the DAC means the gain has to be lowered to around 10, which is in the area where instability may be a problem, the standard gain I have used for the tweeter modules is 15 which is very sweet sounding, but even this is well into clipping territory with 0.5v input.

Alternatively I guess I could attenuate the input to the tweeter amp via fixed resistors and I may yet do this.

I realise this all sounds a bit odd, but trust me the last thing I need is a higher input signal.

Of course I could run it through the active pre-amp, but just 15 seconds of hearing the difference in sound will convince anyone, direct is best.

Glad to hear your DAC is up and running, I will run mine for a few hours tomorrow to see how it sounds with some time on it.
 
Had the dac running through an active pre amp and phones for about 13 hours now.

There is no doubt the sound has smoothed out a lot, far less grain, in fact not really an issue at all except perhaps on complicated material.

The only slight issue is that I feel the sound stage, instrument positioning is a little less precise than with my previous dac, but this might well be just my imagination.

Next, once burnt in I will look at a separate power supply for the dac chip and then the receiver chip. One thing at a time so I can judge the changes.

Does anyone know if there might be a difference in sound by using a set of 3v bias batteries per channel?

If the DAC chip were run from 2500 millamp NiMh cells how long might I expect it to run before recharging?
 
Tolu -

Now do the EC battery mod - it's easy and it makes a big improvement! If you are only using 886 Ohms now you will hardly notice any loss in output.

Zero One -

NiMh cells will give less than 3V, and I think the difference could be significant so they may not work so well.
 
I will try the EC mod!

But there are some questions.

1. How do you calculate the correct V of the battery and the othe depending values (I/V res., Vrms)?

2. In order to get a higher output V I plan to use 4 pcs. of TDA1543 and let them run with 8V. How much V from the battery do I need than?

3. Can anyone put the formula or calculation into the thread?
 
Thomas -
I know someone who tried the EC mod at 8V with various resistor values but had no success. Too much distortion.
I suggest trying the EC values first: 5V supply, 3V battery and 680 Ohms I/V. Disconnect the resistor to pin 7. Single chip.
This is a non-standard way of using this DAC and you may find that no-one has a formula for it.

John
 
A further follow up, I have not yet tried using a pair of 3 v lithiums, ie one for each channel, but will tomorrow, I have something else to sort first.

I took a little time late last night to check the dacs sound out, it was vastly better sounding with 14 hrs or so on it, so clearly there is something to be said for the burning in.

Now about that grain, I could still sense a a bit or it beneath the music and felt I would look harder and tried a few things as follows.

My headphones are very good and the headphone amp utterly silent on battery supplies so it helps when I want to test bit of gear out. So I turned the volume to full with no signal and had a listen, definitely some grainy noise there with some modulating effect in it, mind you this volume level way exceeds anything listenable but a good test.

I realised that the noise was not coming from the DAC but rather the via the trends UD10 transport. I then realised that I was running the transport off the 5v usb supply rather than the very nice SLA regulator supply I have built, hitting the power switchover button killed most of the noise straight away.

Lesson One, that noise coming via the computer even through the Trends UD10 will likely be modulating the musical signal so if using a UD10 or similar build a good supply for it and use it, it does make a difference.

Now when I lisitened again I could here some low level noise/hiss so where was this from?

The Coax cable it seems. If the cable was pulled form the UD10 but still plugged into the DAC the noise remained, no change at all, but pull the coax from the DAC and the noise is totally gone, so it seems clear the cable is picking up some very low level noise. This is not an expensive cable but then not rubbish either. So lesson 2, Two possible things flow from this, a) my system is other than the Mac Mini Battery powered and I have very little in the way of RF pollution in the area or my room, yet the cable is picking stuff up. Consider that most homes will be far worse than mine and since most systems run on mains power my situation probably represents a best case scenario. Could it be that the supposed benefits some folk speak of with better coax cables is not some magic as such, but simple old RF induced effects on the cable, perhaps a better shielded cable may be superior, so that is what I am going to make tonight, make a better cable! Frankly I can't see the point of tweaking bits of the Dac until the fundamentals are nailed.

But wait there's more, moving the USB cable around rendered differing levels of background noise as well, so that means there are also minor issues there.

Finally on top of this neither the USB or Coax are the optimised 1.5 meter length.

It is probably worth mentioning that in the computer world the Mac Mini is pretty much up the top as far as being low noise etc, and I have tried various PCs and Macs, none are as good, so once a gain I hazard that a lot of supposed DAC issues are more likely source issues.

I think it is fair to say the output from the Mac MIni and UD10 transport is probably better in most respects to most reasonably priced CD players so it goes to show that anyone using this or any other dac probably have some things to sort out "source wise" well before they get carried away with changing brands of resistors, caps etc. I am not denying that things things make a difference but I will go out on a limb here and say that in most cases I feel folks are just adding one sort of colouration for another, my feeling is that its better to remove the colouration if possible, hence the total battery powered system (other than the Mac MIni and that will be dealt with eventually)

Next up I checked to see if the sound varies as the dac is lifted away from surfaces, knowing that lots of folk claim the box effects the sound etc. Answer? Actually it did seem to very slightly, raising the DAC 3 or 4 inches in the air off the wooden board it was sitting on seemed to thin out the sound a bit....utterly no idea why, but it means that there is something else to investigate.

Overall I have no doubt this Dac is going to end up sounding very nice indeed, but before I change any bits I am going to sort the cables, then power supplies (separate for all chips), then the shielding of chips etc and then and only then will I try different caps, resistors etc.

Sorry for the long post but I thought the observations may help and it also helps me to focus on the digital pathway at hand.

Thanks for reading.
 
And now for a little follow up.

Having eliminated cable noise, UD10 noise and, checking for power supply noise via batteries making sure no extra caps were needed across the terminals, all was much better except for a small amount of noise, (static mixed with hiss at a very low level and mainly in the left channel, with just a bit of hiss in the right.)

Took a bit of searching but I have found it is the 33K resistor in series with line running between the 74 series input chip and the DIR 9001 receiver chip, touch the resistor and the noise goes completely leaving an utterly silent background even at total max volume through the headphones.

I doubt it is really audible in normal listening....but it has to go.

The question is now how do I eliminate this last noise element?

Any ideas?
 
Zero One said:

The question is now how do I eliminate this last noise element?

Any ideas?


Two possible ways:

1. Keep your finger on the resistor while listening. :D

2. Presumably your body is adding some capacitance when you touch the resistor, so maybe add a small capacitor to ground after the resistor? Have you got a scope that might show what is going on?
 
Thanks Jonners, I was thinking of adding some C though not sure of a value to start with, easy enough to try though. I also considered re soldering the joint in case there is an issue there and perhaps replacing the resistor with an SMD version.

I don't know why it is introducing noise in the first place and I imagine all of the boards would be the same, perhaps it is radiated noise from the nearby 9001 chip, anyway it is worth sorting as it will then leave me with a totally noiseless board which can then be optimised for its audio performance.

Thanks
Zero One
 
I have no sorted all noise issues, running a 1.5 metre quad screened coax, supply for UD10.
better.

So now I have hooked up 2 sets of batteries for the bias, one to each channel.

Strangely it does seem to sound better, less hair on the highs I guess is how I would describe it.
The change however may be due to using Lithiums instead of alkalines so I won't make any great claims but it is sounding very sweet.
 
Zero One said:
I have no sorted all noise issues, running a 1.5 metre quad screened coax, supply for UD10.
better.

So now I have hooked up 2 sets of batteries for the bias, one to each channel.

Strangely it does seem to sound better, less hair on the highs I guess is how I would describe it.
The change however may be due to using Lithiums instead of alkalines so I won't make any great claims but it is sounding very sweet.

Did you exactly the EC mod? With 2x1.5V and 680Ohm I/V and 2.2nF to ground?

Can you post some photos of your modded PCB?

I am very happy with the sound so far! Sweet, I would describe it. Not too fat, no curtain-puller, no detail fetishist, simply pleasant!
 
Hi Tolu

I will take some pics for you soon, I have a few things to do to it yet.

I did mod as per EC, removed the 2200pf caps, made up a good coax using quad screened cable, this is noticeably better than the regular one I was using. The power supply for the tends UD1o transport runs from a 6v SLA regulated with plenty of C (this makes a huge difference in my setup). The current power supply for the DAC is a 7.2 AH SLA, but this will be changed to 4 separate supplies, one for the first chip, 2 for the receiver chip and one for the 1543 chip, all will be battery supplies switched via a set of relays.

2 separate supplies for the bias (lithiums) are used I feel this is better, but it is a close call compared to just one.

The DAC is very much influenced by what is upstream of it, which is why I sorted all that first.

The sound at present, really very nice, actually I find it quite detailed bit not overly so, highs are less bright than other DACs but very tuneable via iTunes with just a little equalization.

Bass is OK but not super strong, mids lovely, what is does to Kate Bushes' voice is really very nice indeed. Overall I could listen to it all day and not tire of the sound but I will seek to take it to a higher level.

The only issue I can identify is a little sibilence and some flabbiness in the bass, but my gut feeling is this is related to the output caps, currently a pair of 3.3 uf greencaps so hardly a top choice but they are what I had on hand.....mind you they are way better than the ones it came with.
 
Sorry Tolu, forgot to mention I replaced the 100 uf cap on the 1543 chip with 1000 uf SMD and a small SMD chip capacitor paralleled to it.

Also the coax is 1.5 meters, which is supposed to be the ideal length to reduce jitter.

I still have to sort a better USB cable to the Trends UD10, it does pick up some noise somehow and is not the optimum lenght.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.