Thanks for the help. I measured the DC offset on my preamp and it was basically unmeasurable. I' ll make the changes after I return from vacation.
My next question regards the "biasing tweak" for the AD8620. I have the UcD180. The boards have a date 06/07/2005. The 3 pads for T21 and T22 have smds mounted. They are both L1VND. What are these devices? Do I already have the "bias tweak"?
Mooglie
My next question regards the "biasing tweak" for the AD8620. I have the UcD180. The boards have a date 06/07/2005. The 3 pads for T21 and T22 have smds mounted. They are both L1VND. What are these devices? Do I already have the "bias tweak"?
Mooglie
Just replaced the opamp decoupling caps on UCD400 AD with these (for older version with ceramic caps). Same footprint as the old ones.
http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=9227849&N=401
First impression is lots of more & refined details but a bit bright. Waiting for them to settle down 🙂
http://uk.farnell.com/jsp/endecaSearch/partDetail.jsp?SKU=9227849&N=401
First impression is lots of more & refined details but a bit bright. Waiting for them to settle down 🙂
Two suggested tweaks to the UCD180AD are:
(a) Bypass the coupling caps between opamp and modulator (output offset at speaker should be ok with AD8620).
(b) Power the AD8620 from a separate 12V supply.
I think these tweaks have actually been proposed by Hypex to further improve the sonics. While the sonics of the UCD180AD are good without modification, a further improvement is always welcome. So why has the user to do these mods, why are the modules not already shipped that way?
The TR100A transformer has separate windings for an auxiliary 12V supply on the HG-module. Its outputs are available at convenient connectors, but there are no corresponding connectors on the UCD!
Instead, to feed the 12V to the AD8620 on the UCD as intended, requires unsoldering two transistors and connecting three supply leads to tiny solder points. Then, you have to drill through the HG-module to cut a ground loop! Sounds scary! I would almost certainly damage the modules. Do I really have to risk this in order to try to use the modules as intended?
Here is what I suggest to Hypex:
- Remove coupling caps on UCD180AD and replace with wire bridges.
- Remove transistors on UCD180AD and provide 12V terminals.
- Remove ground loop on Supply-HG.
This saves cost for Hypex (less components), provides even better sonics, generates even better reviews, and pleases the users. A win-win situation!
Jan-Peter and Bruno, what do you think? Or am I seeing something completely wrong here?
Kurt
(a) Bypass the coupling caps between opamp and modulator (output offset at speaker should be ok with AD8620).
(b) Power the AD8620 from a separate 12V supply.
I think these tweaks have actually been proposed by Hypex to further improve the sonics. While the sonics of the UCD180AD are good without modification, a further improvement is always welcome. So why has the user to do these mods, why are the modules not already shipped that way?
The TR100A transformer has separate windings for an auxiliary 12V supply on the HG-module. Its outputs are available at convenient connectors, but there are no corresponding connectors on the UCD!
Instead, to feed the 12V to the AD8620 on the UCD as intended, requires unsoldering two transistors and connecting three supply leads to tiny solder points. Then, you have to drill through the HG-module to cut a ground loop! Sounds scary! I would almost certainly damage the modules. Do I really have to risk this in order to try to use the modules as intended?
Here is what I suggest to Hypex:
- Remove coupling caps on UCD180AD and replace with wire bridges.
- Remove transistors on UCD180AD and provide 12V terminals.
- Remove ground loop on Supply-HG.
This saves cost for Hypex (less components), provides even better sonics, generates even better reviews, and pleases the users. A win-win situation!
Jan-Peter and Bruno, what do you think? Or am I seeing something completely wrong here?
Kurt
Hmm ... No replies? This thread has been dormant for the past two weeks or so. Vacation time?
Kurt
Kurt
Javin5 said:
So in other words, you'd like them to tweak your module for you? 🙂
The AC coupling caps are there for DC protection, you're not to bypass them unless you know for certain that the offset of your source is low enough that you can get away with it.
The new version of modules do have solder pads to connect aux supplies to. I think the reason for not having it this way is obvious enough. Not everyone has the skill to implement aux supplies, or the money required to do it, or the room for the extra components..... or whatever.
In stock format, the module serves the general market best, in terms of robustness, ease of use, and reliability. Anyone who's willing and able can take them that much further with those simple tweaks.
If they were to follow your recommendations, they'd have all kinds of complaints from all the people who burnt out their speakers from too high an offset, too many emails asking how to implement the auxiliary supply, more complains about the extra cost of the aux supply..
They have to look at the bigger picture obviously, not just what's best for your particular project.
Cheers,
Chris
Chris, thanks for your reply.
It's not a question of having Hypex tweak the modules for me; I simply think that they should ship them as close as possible to their full potential. This may also be to their own advantage, since the market of class D modules is clearly becoming more competitive.
Their UCD-supply HG already has an auxiliary 12V supply; it doesn't need to be built, so I think there should also be corresponding connectors on the UCD180AD. I wouldn't really call this a tweak.
I'm not sure if I understand. Do they now have solder pads or not? If yes, do you know when the change has been made?
I fully agree with you that bridging the coupling caps requires an absolutely dc-free source. If this isn't the case with most preamps (and I simply don't know), then yes, it would indeed be better to ship the UCD180AD with the coupling caps in. Just as a precaution.
Kurt
It's not a question of having Hypex tweak the modules for me; I simply think that they should ship them as close as possible to their full potential. This may also be to their own advantage, since the market of class D modules is clearly becoming more competitive.
Their UCD-supply HG already has an auxiliary 12V supply; it doesn't need to be built, so I think there should also be corresponding connectors on the UCD180AD. I wouldn't really call this a tweak.
I'm not sure if I understand. Do they now have solder pads or not? If yes, do you know when the change has been made?
I fully agree with you that bridging the coupling caps requires an absolutely dc-free source. If this isn't the case with most preamps (and I simply don't know), then yes, it would indeed be better to ship the UCD180AD with the coupling caps in. Just as a precaution.
Kurt
Hi Kurt,
The 400 has solder pads for an aux supply now.
It seems to me you'd still have a few components to remove though, or cut the traces to, if you dare.
It's not everyone that buys the Hypex supply, alot of people like to make their own.
I find the way it's setup up now we get a real decent module at a reasonable price. If it came with every little option to make life easiest for everyone out there, well, someone would still be unhappy, like alot of the people who couldn't afford them anymore.
I kind of agree with you about the parts, but at the same time I appreciate that cost is kept as reasonable as possible. At least this way more people can get their hands on them and later on do the few little tweaks when they can better afford it. After all, not much is kept secret when it comes to what the tweaks are.
Have you gone ahead and implemented the aux supplies btw?
The 400 has solder pads for an aux supply now.
It seems to me you'd still have a few components to remove though, or cut the traces to, if you dare.
It's not everyone that buys the Hypex supply, alot of people like to make their own.
I find the way it's setup up now we get a real decent module at a reasonable price. If it came with every little option to make life easiest for everyone out there, well, someone would still be unhappy, like alot of the people who couldn't afford them anymore.
I kind of agree with you about the parts, but at the same time I appreciate that cost is kept as reasonable as possible. At least this way more people can get their hands on them and later on do the few little tweaks when they can better afford it. After all, not much is kept secret when it comes to what the tweaks are.
Have you gone ahead and implemented the aux supplies btw?
Hi Chris
I have the following parts from Hypex:
- 2 x UCD180AD
- 2 x UCD supply HG
- 2 x transformer TR100A
I built two monoblocs with them, no problem if you carefully check your wiring, they worked the first time. Plenty of power, even with the 180. I'm connecting to my preamp with balanced XLR. The sound is good, but lacking some details.
No, I have not yet tried to connect the aux-supply, although I'm tempted. I think a short while back, there were two specific threads about this, and the conclusion was (if I remember correctly) that, after loooking at the required changes, nobody has actually implemented this and could report about the sonic improvements. I think this tells a lot. So would I be the first one to do this?
There is a diagram for this on the Hypex web site, but it is somewhat unclear. This requires much, much more skills than just bridging the coupling caps. And yes, I'm scared that I would probably damage the modules (600 Euro + shipping + tax).
Kurt
I have the following parts from Hypex:
- 2 x UCD180AD
- 2 x UCD supply HG
- 2 x transformer TR100A
I built two monoblocs with them, no problem if you carefully check your wiring, they worked the first time. Plenty of power, even with the 180. I'm connecting to my preamp with balanced XLR. The sound is good, but lacking some details.
No, I have not yet tried to connect the aux-supply, although I'm tempted. I think a short while back, there were two specific threads about this, and the conclusion was (if I remember correctly) that, after loooking at the required changes, nobody has actually implemented this and could report about the sonic improvements. I think this tells a lot. So would I be the first one to do this?
There is a diagram for this on the Hypex web site, but it is somewhat unclear. This requires much, much more skills than just bridging the coupling caps. And yes, I'm scared that I would probably damage the modules (600 Euro + shipping + tax).
Kurt
Yes, I have. Bridging the coupling caps did somewhat improve the sonics, including the rendering of details. This is at least my impression. But there is still more potential, and, as Hypex pointed out in this forum, feeding the opamp from the aux-supply, would indeed be an additional, significant step forward.
Many on this forum don't use the Hypex supply, they build their own power supply with Jensen or Sicorel caps, etc. It's quite possible, that with such a refined supply, the bridging of the coupling caps would yield even more improvement than with the Hypex Supply-HG.
Here are my measurements of the DC offst voltage at the speakers:
- unmodified: 1 mV / 4 mV
- dc-coupled (coupling caps bridged): 5 mV / 7 mV
Measurements on the two monoblocs were taken after 30 minutes of operation. I really expected somewhat higher offsets after dc-coupling, more in the 20 - 40 mV range. I don't know what is typical over a wide range of modules, but if it is indeed very low, I don't see a point in searching for the best replacement cap; just bridge the thing and you are done. The best cap is no cap. So Hypex might save some componet cost here after all.
By the way, my preamp has of course no measurable DC at the output.
The aux-supply connection to the opamp is still open and on my wish list!
Kurt
Many on this forum don't use the Hypex supply, they build their own power supply with Jensen or Sicorel caps, etc. It's quite possible, that with such a refined supply, the bridging of the coupling caps would yield even more improvement than with the Hypex Supply-HG.
Here are my measurements of the DC offst voltage at the speakers:
- unmodified: 1 mV / 4 mV
- dc-coupled (coupling caps bridged): 5 mV / 7 mV
Measurements on the two monoblocs were taken after 30 minutes of operation. I really expected somewhat higher offsets after dc-coupling, more in the 20 - 40 mV range. I don't know what is typical over a wide range of modules, but if it is indeed very low, I don't see a point in searching for the best replacement cap; just bridge the thing and you are done. The best cap is no cap. So Hypex might save some componet cost here after all.
By the way, my preamp has of course no measurable DC at the output.
The aux-supply connection to the opamp is still open and on my wish list!
Kurt
Javin5 said:
true, if you can assure yourself of zero source offset. Otherwise, the best cap you have becomes a good speaker destroyer and fire hazard.
So this obviously can be done. I still believe that the best opamp is no opamp, unfortunatly there are enough allready in my system that I have to get rid off some day.
Other than that I think it's strange why there is not so much attention to the decoupling of the opamps wich makes a lot of difference on opamp performance. I have had good very good results with 100uF/16V BG-NX in super e-cap configuration decoupled with small coils in the past with my CD player, took a lot of stressssss out of the sound.
What the gain of the input opamp stage. I thought I have read somewhere it's only a gainstages of a factor 2, that wouldn't be to hard to overcome with a decent preamp.
Gain factor?
The kind of input circuit used requires a small but precisely sourced current. To source this current into the input resistors requires a nearly perfect source voltage of close to “0” ohms impedance. Point I am making is that with this type of configuration the source impedance is critical and must be balanced. A typical preamp with 100’s if not 1000’s of ohms source impedance will cause a small to severe distortion and gain loss as well as adding noise. It isn’t so bad if the +/- sources are balanced at all audio frequencies and even better if balanced at Hf.
These types of problems are why the buffer op amp is there to begin with. It is configured to maximize common mode rejection, supply modest gain and most important to transform a very high input impedance to a very low output impedance. I am not saying your idea can’t work successfully but there are a bunch of factors to be properly considered.
Roger
The kind of input circuit used requires a small but precisely sourced current. To source this current into the input resistors requires a nearly perfect source voltage of close to “0” ohms impedance. Point I am making is that with this type of configuration the source impedance is critical and must be balanced. A typical preamp with 100’s if not 1000’s of ohms source impedance will cause a small to severe distortion and gain loss as well as adding noise. It isn’t so bad if the +/- sources are balanced at all audio frequencies and even better if balanced at Hf.
These types of problems are why the buffer op amp is there to begin with. It is configured to maximize common mode rejection, supply modest gain and most important to transform a very high input impedance to a very low output impedance. I am not saying your idea can’t work successfully but there are a bunch of factors to be properly considered.
Roger
Sjef said:
Hi Sjef,
I think you are using the DEQX PDC? If that is the case, you have a while bunch of OPA2134 in the signal path. The UcD comes standard with the OPA2134, so one more is probably not going to make much of a difference.
Maybe, the PDC balanced output could be directly connected to the UcD modulator to get at least one OPA2134 out of the signal path. Don`t know if the PDC can give enough output signal to drive the modulator directly (the opamp can obviously do it).
Best regards
Gertjan
Anyone tried this little mod to the power supply which i`m using which improves performance : a secondary raw supply for the regulated section.As I had some spare T-Network caps and schottky rectifiers I thought I would wire these in parallel with the main supply and feed the outputs to the pair of voltage dropping resistors, which need to be lifted at one end, in my case a no cost improvement!
Bob Lewis
Bob Lewis
Yes Geertjan, I'm using the DEQX as preamp/crossover same as you do, more people should do that actually. After three years of intensive listening to the DEQX all I can tell it's a great unit when used correctly. That is as active crossover with all the correction done etc. This can make a very big difference in every area. As stand allone preamp my former tube amp performs better in sence of detail, dynamics, air etc. As standallone preamp the sonic signature of the opamps becomes audible. Can only dream of a tubed DEQX.
On a dutch forum Jan-peter has stated that the UCD700 sounds better than the 400 and the 180 probably because they needed to feed the buffer from a seperate power supply. That's also my experience with opamps, they are very sensible to the power supply used, seperate power supply is really needed to make the best out of them.
On a dutch forum Jan-peter has stated that the UCD700 sounds better than the 400 and the 180 probably because they needed to feed the buffer from a seperate power supply. That's also my experience with opamps, they are very sensible to the power supply used, seperate power supply is really needed to make the best out of them.
Sjef said:
Hi Sjef,
I may try seperate supply for the opamps sometime, once I have enough seperate supplies. I basically want to go to an all monoblock configuration, meaning thus 6 amps and 6 supplies. This is costly but it allows me to have seperate grounds for all amps so that for example the tweeter amp GND is not influenced by high currenst drawn by the woofer amp and/or midrange amp. After cleaning up the GND level that way, I will think about doing seperate opamp supply. I could easily do this when I use the coldamp SMPS (SPS80 and or SPS30 (not yet out)).
I think it should be possible to reduce the number of opamps in the PDC, however, then I would like to have schematics of the PDC. It seems the PDC has DC servo controlled outputs (to remove DC offset, so no output caps needed), so some of the opamps should be for that purpose. I guess there maybe still 3 opamps or so in the signal path.
If you find a good way to reduce the number of opamps in the PDC signal path, let me know 🙂
Best regards
Gertjan
LM4562
The LM4562 op amp is now on the National site:
http://cache.national.com/ds/LM/LM4562.pdf
This looks to be a very good possibility for UcD module upgrades.
About the only thing I see not ideal is that without changing the supply rails its swing is a bit limited.
But .00003 % THD +N and >100dB PSSR and CMMR are going to be positive.
Its a bipolar input, but with 10nA, it shouldn't be to much offset.
Mike
The LM4562 op amp is now on the National site:
http://cache.national.com/ds/LM/LM4562.pdf
This looks to be a very good possibility for UcD module upgrades.
About the only thing I see not ideal is that without changing the supply rails its swing is a bit limited.
But .00003 % THD +N and >100dB PSSR and CMMR are going to be positive.
Its a bipolar input, but with 10nA, it shouldn't be to much offset.
Mike
- Status
- Not open for further replies.