Hi Davide,
You can easily push the rails up a little bit, but performance actually starts to degrade above 50V or so. Current also increases at higher voltages, so the regulator dissipation won't drop as much as you might be thinking. At 50V rails, the current per reg will be 130mA instead of the 115mA you'd see at 45V. Just the same, it is a valid way of transferring some of the load off the regs and back to the main heatsink.
With the transformers you have (40VAC), you'll be looking at about 55VDC before the regs, which means you're dissipating about 1.2W per regulator. That should be an acceptable amount, but depending on the transformers regulation and your AC line level, it could be higher. If you're around 60V before the regulators, then it's too high and you either need to swap transformers, or drop a few volts by adding some resistance to the secondary (not an elegant solution).
Regards,
Owen
You can easily push the rails up a little bit, but performance actually starts to degrade above 50V or so. Current also increases at higher voltages, so the regulator dissipation won't drop as much as you might be thinking. At 50V rails, the current per reg will be 130mA instead of the 115mA you'd see at 45V. Just the same, it is a valid way of transferring some of the load off the regs and back to the main heatsink.
With the transformers you have (40VAC), you'll be looking at about 55VDC before the regs, which means you're dissipating about 1.2W per regulator. That should be an acceptable amount, but depending on the transformers regulation and your AC line level, it could be higher. If you're around 60V before the regulators, then it's too high and you either need to swap transformers, or drop a few volts by adding some resistance to the secondary (not an elegant solution).
Regards,
Owen
definitely make sure the trimpots are turned all the way up and turn down to adjust (without any dac connected of course), they wont allow it to go too high with the stock resistors anyway from memory, this will help to avoid blowing the regs with too much drop across them.
also about the input pin bias, actually it should be higher than 1.65, the tp guys set the avcc shunt higher than 3v3 for high dnr as we discussed, so you should measure the actual dc output of your avcc regs and divide that by 2.
ive found Richards transformers to have excellent regulation, so you should be pretty safe there.
also about the input pin bias, actually it should be higher than 1.65, the tp guys set the avcc shunt higher than 3v3 for high dnr as we discussed, so you should measure the actual dc output of your avcc regs and divide that by 2.
ive found Richards transformers to have excellent regulation, so you should be pretty safe there.
Last edited:
I guess it's time for me to understand how this toy works:
It kind of play music, but something does not work as it should. Let's say that I start with 0V on the DAC pins. It works. Now I start increasing the voltage. it looks like it increase linearly, up to 0.8V then it does not go further and if I continue turning the mosfet shut off and I have all of a sudden almost the same voltage on drain and source.
Looking also at the distortion and amplitudes, it looks like I get the best performances with the pin voltages at 0.390 V.
Why do you think it does not behave as it should ?
The supply voltages are all 45 V.
Thanks,
D.
It kind of play music, but something does not work as it should. Let's say that I start with 0V on the DAC pins. It works. Now I start increasing the voltage. it looks like it increase linearly, up to 0.8V then it does not go further and if I continue turning the mosfet shut off and I have all of a sudden almost the same voltage on drain and source.
Looking also at the distortion and amplitudes, it looks like I get the best performances with the pin voltages at 0.390 V.
Why do you think it does not behave as it should ?
The supply voltages are all 45 V.
Thanks,
D.
I'm going to bet you have the protection diodes installed at the inputs, and you didn't implement the fix that's required there.
There is an error on the board where I put the input protection zener diodes in parallel, rather than in series. If you look at the schematic on the first page of this thread, you'll see that Z4, Z2, Z6, Z8 as well as Z1, Z3, Z7, and Z5 are each in parallel at the input when they need to be in series. this is mentioned a few times in the build thread, but it's easy to miss.
So... when you go to adjust the input voltage up to 1.65V, both Z2 and Z6 will start conducting at about 0.7-0.8V and essentially shunt your drains to ground. That's why you can't get above 0.8V, and also why things aren't working properly.
You need to remove all those diodes and do one of two things:
1. Leave them unpopulated - I did this and all is well.
2. re-populate them but solder them in series, anodes facing outward, by making a little teepee with both diodes where one used to be. I've attached a picture from qusp for clarity.
Implement that fix and you should be ready to go!
Regards,
Owen
There is an error on the board where I put the input protection zener diodes in parallel, rather than in series. If you look at the schematic on the first page of this thread, you'll see that Z4, Z2, Z6, Z8 as well as Z1, Z3, Z7, and Z5 are each in parallel at the input when they need to be in series. this is mentioned a few times in the build thread, but it's easy to miss.
So... when you go to adjust the input voltage up to 1.65V, both Z2 and Z6 will start conducting at about 0.7-0.8V and essentially shunt your drains to ground. That's why you can't get above 0.8V, and also why things aren't working properly.
You need to remove all those diodes and do one of two things:
1. Leave them unpopulated - I did this and all is well.
2. re-populate them but solder them in series, anodes facing outward, by making a little teepee with both diodes where one used to be. I've attached a picture from qusp for clarity.
Implement that fix and you should be ready to go!
Regards,
Owen
Attachments
That's why I love this forum, everything is back on track.
Uncle Tom's cabin with diodes was far beyond my smd soldering capabilities, so I left the diodes not populated.
Thanks,
I'll let it run for a bit more and then move to the listening room.
I'm still a bit worried about the final dissipation. It runs kind of hot. My transformers have 13 % regulation, now I am running at 46 V. Not a big change.
Aren't zener with higher voltage available that can fit there ?
Thanks,
D.
Uncle Tom's cabin with diodes was far beyond my smd soldering capabilities, so I left the diodes not populated.
Thanks,
I'll let it run for a bit more and then move to the listening room.
I'm still a bit worried about the final dissipation. It runs kind of hot. My transformers have 13 % regulation, now I am running at 46 V. Not a big change.
Aren't zener with higher voltage available that can fit there ?
Thanks,
D.
Hi Davide,
The zener voltage is not relevant when the parts are forward biased as they would be in this setup. The zener voltage only applies when they're reverse biased.
When they're place in series, they won't start to conduct until you exceed the zener voltage of the first diode plus the forward voltage of the second diode. So if you have a pair of 3V zeners, then the whole arrangement will conduct when you exceed 3.7V. With the way they are now on the board, you could use 100V zeners and it would still not allow you to get above 0.7V.
The real solution would be to use a part like this:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=568-4044-1-ND
That part is basically two series zeners in one pack that would fit on the existing footprint.
Did you measure your DC voltage before the regs? Could you let me know what it is?
Regards,
Owen
The zener voltage is not relevant when the parts are forward biased as they would be in this setup. The zener voltage only applies when they're reverse biased.
When they're place in series, they won't start to conduct until you exceed the zener voltage of the first diode plus the forward voltage of the second diode. So if you have a pair of 3V zeners, then the whole arrangement will conduct when you exceed 3.7V. With the way they are now on the board, you could use 100V zeners and it would still not allow you to get above 0.7V.
The real solution would be to use a part like this:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=568-4044-1-ND
That part is basically two series zeners in one pack that would fit on the existing footprint.
Did you measure your DC voltage before the regs? Could you let me know what it is?
Regards,
Owen
I have 56.5 V on the cap before the regulators. From the spreadsheet this put 2.5 W on each regulator and almost 44 W on the heatsink.
If I get to 50 V I will have 1.7 W on the little sinks and 52 W on the big heatsink.
What do you advice ?
Thanks,
D.
If I get to 50 V I will have 1.7 W on the little sinks and 52 W on the big heatsink.
What do you advice ?
Thanks,
D.
At the moment, with the heatsink facing up I have 55 degrees on the big one and 65 degrees on the little one. If I turn it the other way (board up sink down) I have 58 degrees on the big heatsink.
It'a difficult to assess how it will be in the cabinet, as some of the heat will be transferred to the base plate, plus the heatsink will be a bit elevated, with free air circulation underneath. and from side to side. But the top plate will limit a bit the dissipation.
So I think I just have to wait to receive the plates and do some measurement.
By the way. The temperatures are taken with an IR thermometer with epsilon set to 1.00. Don't know how accurate it is, and the room temperature is 27 degrees.
Thanks,
D.
It'a difficult to assess how it will be in the cabinet, as some of the heat will be transferred to the base plate, plus the heatsink will be a bit elevated, with free air circulation underneath. and from side to side. But the top plate will limit a bit the dissipation.
So I think I just have to wait to receive the plates and do some measurement.
By the way. The temperatures are taken with an IR thermometer with epsilon set to 1.00. Don't know how accurate it is, and the room temperature is 27 degrees.
Thanks,
D.
Hi Davide,
Those numbers seem reasonable to me. You'll need to keep an eye on things when you mount it in a case, but the setup you described should help rather than hinder.
Try measuring the ESS DAC itself and see what temp it's sitting at. I'd bet it's reasonably low.
In a perfect world, you'd be best to get a new pair of 35VAC transformers and run the I/V at 45V. That would alleviate the dissipation on the regs and keep the heatsink temp a little lower.
In a non perfect world, you're probably better off transferring some of the load to the larger sink so that the regulators run a little cooler. Once you mount it in a case, most of the heat from the large sink will stay outside the case, where the heat from the regs will be captive inside the case. On that note you'd be best to minimize the load on the regs.
Regards,
Owen
Those numbers seem reasonable to me. You'll need to keep an eye on things when you mount it in a case, but the setup you described should help rather than hinder.
Try measuring the ESS DAC itself and see what temp it's sitting at. I'd bet it's reasonably low.
In a perfect world, you'd be best to get a new pair of 35VAC transformers and run the I/V at 45V. That would alleviate the dissipation on the regs and keep the heatsink temp a little lower.
In a non perfect world, you're probably better off transferring some of the load to the larger sink so that the regulators run a little cooler. Once you mount it in a case, most of the heat from the large sink will stay outside the case, where the heat from the regs will be captive inside the case. On that note you'd be best to minimize the load on the regs.
Regards,
Owen
Actually for the time being I could use the 115 V tap with 100 V main, This should solve the problem. When I will be back to Europe next year I can see what to do.
D.
D.
That would work perfectly. Even if you have to drop down to 40v rails, there's very little difference between 40V and 45V. You're best to minimize the drop across the regs if you think things are running a little too hot.
Give it a try and see what you get before the regs, then just set the output of the regs a few volts lower.
Regards,
Owen
Give it a try and see what you get before the regs, then just set the output of the regs a few volts lower.
Regards,
Owen
and we have another winner!!
yeah Davide, the numbers i did on the sink and my personal experience put your 27 degree ambient temp in the upper range, but still within SOC as long as you think about it wrt the other stuff in the case, ventilation etc.
yeah Davide, the numbers i did on the sink and my personal experience put your 27 degree ambient temp in the upper range, but still within SOC as long as you think about it wrt the other stuff in the case, ventilation etc.
Last edited:
Swapping the taps I have 49.5 V DC in and I set 45Vdc out. That's just perfect. 1 W on the little sink is ok.
I could listen to it on my main system and made some measurements. Well, with my instruments, measurement were boring, there was just nothing. But I found out that I have a problem on the DAC at 192 kHz. I will have a new usb-i2s interface soon and I will compare.
From the listening point of view. Compared to the legato, it's another world, beside the fact that the input is much higher, and I do not run out of juice with my F5 on the audiophile recordings, that are recorder lower then normal rock albums.
I don't think I will feel the need to go dual mono, but I will give it a try.
Thanks,
Davide
I could listen to it on my main system and made some measurements. Well, with my instruments, measurement were boring, there was just nothing. But I found out that I have a problem on the DAC at 192 kHz. I will have a new usb-i2s interface soon and I will compare.
From the listening point of view. Compared to the legato, it's another world, beside the fact that the input is much higher, and I do not run out of juice with my F5 on the audiophile recordings, that are recorder lower then normal rock albums.
I don't think I will feel the need to go dual mono, but I will give it a try.
Thanks,
Davide
Hi Davide,
That's great news all around! I didn't realize it was so easy for you to swap taps on the primary, or I would have suggested it earlier.
50V input is pretty much perfect for 45V rails, so you're right where you need to be.
I'm glad you like the way it sounds, you've certainly earned it! It's a really kick-*** I/V stage, and I always enjoy it when people like the way it sounds.
Good work on the assembly, and I'm glad you hung in there through the little ups and downs of getting it just right.
Enjoy 🙂
Owen
That's great news all around! I didn't realize it was so easy for you to swap taps on the primary, or I would have suggested it earlier.
50V input is pretty much perfect for 45V rails, so you're right where you need to be.
I'm glad you like the way it sounds, you've certainly earned it! It's a really kick-*** I/V stage, and I always enjoy it when people like the way it sounds.
Good work on the assembly, and I'm glad you hung in there through the little ups and downs of getting it just right.
Enjoy 🙂
Owen
Yes, I'm kind of postponing the problem, but in Europe is pretty easy to get toroids, don't ask me why Japanese don't use them, at least you cannot buy them easily.
I still did not decide what to do with the balanced to unbalanced stage. I have one board nicely boxed that I am using now, and a full kit ready to be assembled.
Basically I have three options:
1) Stay as I am. Hopefully the F5X will arrive soon and I will not need it. When I need it I can use the external unit.
2) Stick the converter in the DAC, in order to have SE out.
3) Stick the converter in the preamp (DCB1) in order to have a balanced input in the pre.
Maybe the latter is the most versatile option. Additionally this would leave me more space in the chassis if I want to experiment with the power supplies of the DAC or of the IV.
I think it was my mistake with the transformers, as I ordered the 50 VA, with the idea bigger is better, but with small current drain the voltage is too high. I should have told Richard to use the bigger core and reduce the regulation. Lesson learned.
For sure this is not a beginner project, both for the SMD part and for the manually needed to assemble the board on the sink.
I ordered the diodes you suggested, but what's the real risk, once the thing is fully assembled and tuned ?
Thanks,
Davide
I still did not decide what to do with the balanced to unbalanced stage. I have one board nicely boxed that I am using now, and a full kit ready to be assembled.
Basically I have three options:
1) Stay as I am. Hopefully the F5X will arrive soon and I will not need it. When I need it I can use the external unit.
2) Stick the converter in the DAC, in order to have SE out.
3) Stick the converter in the preamp (DCB1) in order to have a balanced input in the pre.
Maybe the latter is the most versatile option. Additionally this would leave me more space in the chassis if I want to experiment with the power supplies of the DAC or of the IV.
I think it was my mistake with the transformers, as I ordered the 50 VA, with the idea bigger is better, but with small current drain the voltage is too high. I should have told Richard to use the bigger core and reduce the regulation. Lesson learned.
For sure this is not a beginner project, both for the SMD part and for the manually needed to assemble the board on the sink.
I ordered the diodes you suggested, but what's the real risk, once the thing is fully assembled and tuned ?
Thanks,
Davide
Just one question: if I want to use the spdif isolation transformer for normal spdif in the buffalo input, I think the resistor should not be populated at all, as the termination is already on the board. Is it correct ?
Thanks,
Davide
Thanks,
Davide
no the resistor on the board iv should be populated and just turn the spdif switch on the b2 to off and use D1 and ground rather than spdif and ground on the buffalo. afaik its best to have termination at the end of the cable before the transformer, otherwise the sending end will only see the reflected impedance. I could be wrong there, but thats the way i understand it
I'd like to avoid to use D1, so I can devote that to I2S and switch the input via software.
The transformer is 1:1 so should not matter on which side you put it, but in all the reference design is towards the receiving device.
D.
The transformer is 1:1 so should not matter on which side you put it, but in all the reference design is towards the receiving device.
D.
haha sorry you are right, my utterly non-photgraphic memory is faulty. for some reason i was thinking the resistor was at the input of the schematics i was looking at last night in order to terminate the cable. my comment still holds re D1, but if you have other plans for that pin then sure, if you have to keep the comparator and termination on the buff in place then dont populate the one on the board.
Davide,
Could you elaborate a bit on your listening experience? Different or better?
Cheers,
Nic