Quick question for you guys. I was wondering if anyone knows if there are any DAC chips with the same (or similar) pinout / size as the Burr Brown PCM58P?
I've been looking around for a while now, but the chip seems old enough that most of the information has disappeared.
I'd like to replace this chip in a piece of equipment and perhaps get something more modern, perhaps up to 24/192.
Is this something anyone has done or am I being a dumbass?
I've been looking around for a while now, but the chip seems old enough that most of the information has disappeared.
I'd like to replace this chip in a piece of equipment and perhaps get something more modern, perhaps up to 24/192.
Is this something anyone has done or am I being a dumbass?
I have the data sheet of PCM56P which I believe has the same pinout as the PCM60P. Great sounding DAC chip.
That would be beautiful. Do you know what input formats the 60P will accept by chance?
I attached the data sheet of PCM56P. I never played with it except than I have a Discman with a PCM60P inside...
Stop and think. All the datasheets are readily available. I downloaded the PCM58 specs and learned it is an 18-bit, monaural, DAC. It came in a 28-pin DIP package and required +5v and -12v power. It was discontinued in 1997. I doubt you will find a pin compatible replacement for it. (Why would you want to?) Neither the PCM56 or PCM60 are compatible because both are 16-bit devices with different packaging, pin-outs, and power requirements.
The 20-bit PCM63 is the logical successor to the PCM58 but it has different pin-outs and voltage requirements.
As for 24/192, the PCM58 can already do that.
the PCM58 was a nice dac, used in some well-regarded machines. there is no drop-in upgrade or replacement. pcm63 was next one up the food chain, but if you do the rework to use that one, you might be better off just changing for PC1702 or PCM1704.
;-)
mlloyd1
;-)
mlloyd1
Ok, it seems like my grasp on how dacs work is a little lackluster.
Time to do a little more research.
Resurrection....
I have an old Aragon D2A with a pair of these chips. I just replaced the electrolytic on the digital board and also the mystery coupling caps on the analogue board. These boards are crap, and don't like rework by the way... Many of the otiginal traces are lifted from original assembly!
Anyway, I was just looking at the chip data sheet, and it has provision for external bit correction. Aragon apparently took advantage of the first bit correction option, as shown by the trim pot next to the chip. But, there are also provisions for the 2nd, 3rd and 4th bit to be corrected, for even lower THD+noise. It seems easy to implement, based on my investigations.
Has anyone tried this? The relevant pins of the chip are (to be) left open when not making this adjustment, so in theory, they are currently open and you could "tap into" the chip on the back side of the board, and then run a daughter board with the three additional pots on it to set zero.
So, that's the mechanical side of it, which I understand. The actual adjustment is another story... A sine wave input and the adjustment made with a scope(?) to determine lowest THD+Noise. Not hard for someone that knows what their doing, but I'm not that guy...
"Bits 2 through 4 can also be adjusted if desired to obtain optimum full-scale output THD+N performance. An additional 100kΩ potentiometer adjustment circuit is required for every additional bit to be adjusted. If bit adjustment is not performed, the respective pins on the PCM58P should be left open." (See page 5 of the attached)
I have an old Aragon D2A with a pair of these chips. I just replaced the electrolytic on the digital board and also the mystery coupling caps on the analogue board. These boards are crap, and don't like rework by the way... Many of the otiginal traces are lifted from original assembly!
Anyway, I was just looking at the chip data sheet, and it has provision for external bit correction. Aragon apparently took advantage of the first bit correction option, as shown by the trim pot next to the chip. But, there are also provisions for the 2nd, 3rd and 4th bit to be corrected, for even lower THD+noise. It seems easy to implement, based on my investigations.
Has anyone tried this? The relevant pins of the chip are (to be) left open when not making this adjustment, so in theory, they are currently open and you could "tap into" the chip on the back side of the board, and then run a daughter board with the three additional pots on it to set zero.
So, that's the mechanical side of it, which I understand. The actual adjustment is another story... A sine wave input and the adjustment made with a scope(?) to determine lowest THD+Noise. Not hard for someone that knows what their doing, but I'm not that guy...
"Bits 2 through 4 can also be adjusted if desired to obtain optimum full-scale output THD+N performance. An additional 100kΩ potentiometer adjustment circuit is required for every additional bit to be adjusted. If bit adjustment is not performed, the respective pins on the PCM58P should be left open." (See page 5 of the attached)
Good grief, that was obsolete when I designed my DAC 25 years ago. I used a PCM1702-K and suggest you do too. An adapter board will be required.
I’m not designing around it- I’m looking into a potential improvement of it as it is.
I wasn’t aware of the trim pot’s function until I found that sheet. So now I’m wondering how much more can be wrung from these “antique” chips?
Further, perhaps the analogue section can be improved with some modern chips…it’s predominantly discrete but some transistors as well.
Why bother you ask? Why not I say!
I wasn’t aware of the trim pot’s function until I found that sheet. So now I’m wondering how much more can be wrung from these “antique” chips?
Further, perhaps the analogue section can be improved with some modern chips…it’s predominantly discrete but some transistors as well.
Why bother you ask? Why not I say!
I'm not at all certain that this will work, but for trimming, you could use input signals consisting of a very small sinewave added to an offset, where the offset is chosen such that the bit you want to trim and all the lower bits switch.
For example, for trimming the second most significant bit, you would need an offset that brings it to either one quarter or three quarters of its range (I mean of the entire range from most negative to most positive code). In offset binary code that's 010000... or 110000..., in two's complement 110000... or 010000...
For the bit below that, it would be some odd multiple of an eighth, so 011000... or 101000... in offset binary / 111000... or 001000... in two's complement should do.
For the MSB, you don't want any offset, so 100000... in offset binary or 000000... in two's complement.
Audio editors usually have a range from -1 to +1, so in an audio editor, the required offset would be 0 for the MSB, plus or minus 0.5 for the bit below the MSB, plus or minus 0.25 for the bit below that and so on. It would be simple to generate suitable .wav files using the GoldWave expression evaluator.
As the bit settings influence each other, you would have to start from the lowest bit you want to trim and work upwards.
For example, for trimming the second most significant bit, you would need an offset that brings it to either one quarter or three quarters of its range (I mean of the entire range from most negative to most positive code). In offset binary code that's 010000... or 110000..., in two's complement 110000... or 010000...
For the bit below that, it would be some odd multiple of an eighth, so 011000... or 101000... in offset binary / 111000... or 001000... in two's complement should do.
For the MSB, you don't want any offset, so 100000... in offset binary or 000000... in two's complement.
Audio editors usually have a range from -1 to +1, so in an audio editor, the required offset would be 0 for the MSB, plus or minus 0.5 for the bit below the MSB, plus or minus 0.25 for the bit below that and so on. It would be simple to generate suitable .wav files using the GoldWave expression evaluator.
As the bit settings influence each other, you would have to start from the lowest bit you want to trim and work upwards.
Last edited:
i think you need equipment to properly set those pots. they descibe the approach in the data sheets. if you've got the equipment to do it properly and don't mind spending the time, try it and see what you think. then come back and let us know. i used to have a player that used this DAC years ago. it sounded pretty good. it did not have those DAC adjustments on the PCB, but the machine still measured very well, especially for that time.
They only explicitly describe the procedure for the MSB: adjust for minimum distortion with a low-level sine wave and no offset.
I don’t fully understand the procedure, but I see the potential there.
Does anyone have a high-end player or converter with these chips where there are four trim pots per chip?
ESL is close to my work so I could stop in and chat with them about it. If they can properly set it up, I can make the daughter board and deliver it all to them.
Does anyone have a high-end player or converter with these chips where there are four trim pots per chip?
ESL is close to my work so I could stop in and chat with them about it. If they can properly set it up, I can make the daughter board and deliver it all to them.
You maybe overstate the importance of the dac chip in the whole process of what you ear from your dac device. Those chips indeed need trimming as per what is writted in their datasheets and so you need a scope and knowledge.
I use such chips in dac projects without using the trimming function and it is still good enough also with better chips like the AD1862 dac chips. For illustration you can read the AD1862 thread from Miro1360 member, same PCM dac chip architecture of the old days, still very good at ears. In the project most members use it without trimming and let this stage unpopulated.
It's true if you don't use trimming, K grade chips (sorted out) will give you better result.
I subejectivly prefer the PCM56, but the 58 is still a very good chip while you don't really care the two more bits for listening in real world, most of our material are 16 bits.
If you want to improve this dac device you should better focus on the I/V and buffer output stage for better modern operational amplifier and let like it is the pots that were setuped by the brand at the factory. You also can put the dac on a shelf and keep it as a sounding reference for benchmarking and build for few a better sounding dac here : DAC AD1862: Almost THT, I2S input, NOS, R-2R , pcb design is free. Oh and the author made also a pcb for your dac chips and sibblings: read the first post.
Just my two cents, YMMV.
I use such chips in dac projects without using the trimming function and it is still good enough also with better chips like the AD1862 dac chips. For illustration you can read the AD1862 thread from Miro1360 member, same PCM dac chip architecture of the old days, still very good at ears. In the project most members use it without trimming and let this stage unpopulated.
It's true if you don't use trimming, K grade chips (sorted out) will give you better result.
I subejectivly prefer the PCM56, but the 58 is still a very good chip while you don't really care the two more bits for listening in real world, most of our material are 16 bits.
If you want to improve this dac device you should better focus on the I/V and buffer output stage for better modern operational amplifier and let like it is the pots that were setuped by the brand at the factory. You also can put the dac on a shelf and keep it as a sounding reference for benchmarking and build for few a better sounding dac here : DAC AD1862: Almost THT, I2S input, NOS, R-2R , pcb design is free. Oh and the author made also a pcb for your dac chips and sibblings: read the first post.
Just my two cents, YMMV.
Last edited:
In the Aragon (what I'm wortking with) the digital board is seperate from the analogue board. I had the chips on the difgital board reviewed by a Ph.D. who, in a former life, designed the chips themselves. He found no glaring errors or massive room for improvement on the digital board.
I, proactively perhaps, replaced the one electrolytic on the digital with a new one of higher voltage rating.
The analogue board, is fully discrete. Perhaps there are better transistors available that offer a lower noise floor than what is currently installed. That is possible, and something I have not yet investigated. I have replaced the "mystery" coupling caps with new metallized film caps of equal value but much tighter tolerance (Kemet 5%). Otherwise, I'm leaving it alone inside. I purchased replacement electrolytics for it (12 of them) but due to the very poor quality of the board, I elected not to tackle that quite yet...
I, proactively perhaps, replaced the one electrolytic on the digital with a new one of higher voltage rating.
The analogue board, is fully discrete. Perhaps there are better transistors available that offer a lower noise floor than what is currently installed. That is possible, and something I have not yet investigated. I have replaced the "mystery" coupling caps with new metallized film caps of equal value but much tighter tolerance (Kemet 5%). Otherwise, I'm leaving it alone inside. I purchased replacement electrolytics for it (12 of them) but due to the very poor quality of the board, I elected not to tackle that quite yet...
drop a photograph !
Notice than different caps change the hearing result. But it's not an automatical improvement... it can also be worse !? In fact it asks experience in the caps choice. The lythics should be good enough yet. Better to keeep your monney and fresh caps for a better new cheap diy design or instead a cheap modern DAC à la Topping30 or Ifi to test if you like DeltaSigma better or not than your PCM DAC. But for sure the PCM58 dac chip is not the culprit, whatever there are indeed even better PCM dac chip. Anyhow the PCM58 was one of the good PCM chips.
Notice than different caps change the hearing result. But it's not an automatical improvement... it can also be worse !? In fact it asks experience in the caps choice. The lythics should be good enough yet. Better to keeep your monney and fresh caps for a better new cheap diy design or instead a cheap modern DAC à la Topping30 or Ifi to test if you like DeltaSigma better or not than your PCM DAC. But for sure the PCM58 dac chip is not the culprit, whatever there are indeed even better PCM dac chip. Anyhow the PCM58 was one of the good PCM chips.
Bass response has always been an area of concern with these units, and those original coupling caps have been replaced in both this D2A as well as the 47k. They were not very high grade from new (electrolytics were Sprague) and I wasn’t inclined to go for a $100 each “boutique” cap.
Quality metallized film with tighter tolerance should do it after a good break-in period.
Meanwhile I’m pushing forward with the added bit trim pot mod.
Quality metallized film with tighter tolerance should do it after a good break-in period.
Meanwhile I’m pushing forward with the added bit trim pot mod.
That's easy to do. Cut the legs of the old caps close to the body leaving the wire stumps soldered to the PCB. Trim and solder the legs of the new caps to the stumps.
As for the rest, I won't try to talk you out of anything because you seem to be keen on polishing a turd. But I'll ask you, if simply adding a few trim pots would turn a sow's ear into a silk purse, why didn't the hot shot designers of the day do it? Could it be they realized an 18-bit DAC, no matter how it was hot-rodded, could not compete with the new 20- and 24-bit sign-magnitude DACs. Or you could read the datasheet and learn the trimming was added to make the DAC more attractive for use in instrumentation.