Hotrodding the UCD modules

Status
Not open for further replies.
Jump to Latest
#181
UCD 180AD

This maybe interesting or not but you tell me:

Was looking at my modules planning out my tweaks when I decided to get my magnet out. Wanted to check the input & output pins to see if they we ferrous metal. Nope passed magnet test.

Ran magnet over bottom of module and felt a little tug. Went back and sure enough was there. I unscrewed one of the two small bolts holding the T heatsink and sure enough it stuck to the magnet like a tick stuck to a dog. Except these two bolts, rest of the bottom was ok.

Decided to check the rest of the top side of the module. Small bolt, 3 washers, & nut securing the voltage regulators to the T sink were also all sticking to the magnet.

The only other area that was attracted to the magnet was the metal band that surrounds the coil housing. The number on this part reads EP-17 30uH 180W 0426K.

Dont know if this matters or not, it bothers me. I will definately replace the above mentioned bolts, washers, & nuts with brass, or plastic, or stainless steel depending on what I can find.
 
#182
Aslo of note. I posted over at Audio Asylum tweak section looking for advice for a replacement for the 0.68uf 63v cap.

John at Audience posted this response:

"Audience is designing an Auricap capacitor for the application in the Hypex UCD digital amp that you are inquiring about. Feel free to check with me in a few weeks regarind the time of the release of the product. John McDonald" 😉
 
#183
New Auricap

Stevenacnj said:
Aslo of note. I posted over at Audio Asylum tweak section looking for advice for a replacement for the 0.68uf 63v cap.

John at Audience posted this response:

"Audience is designing an Auricap capacitor for the application in the Hypex UCD digital amp that you are inquiring about. Feel free to check with me in a few weeks regarding the time of the release of the product. John McDonald" 😉
Click to expand...

Unfortunately, this announcement is premature. Only the preliminary design work has been done and it is not known if it can be produced yet, that is why I never said anything about it. Yes, I was the one who put the bug in their ear about this and got the ball rolling. It could be a very significant part upgrade if it can be produced. I am following up on it and will post anything significant. The planed part will be a PC mount type and a direct drop in replacement.
Roger
 
#184
IVX said:

Bruno,
how much THD came from the gaped EP17 and RM10 (UcD400?), at the rated current and 1khz?
Click to expand...

Depends what you mean by rated current. I design the coil for an inductance drop of 10% at the protection limit (9..10A for UcD180, 20A for UcD400). At that current, clearly distortion performance of the output filter isn't brilliant (1%). 3dB lower we're on the order of 0.01% which is below the self-alias distortion of the amp. THD is 3rd, so it drops further 40dB per 20dB level decrease.
 
#186
Originally posted by Bruno Putzeys

I design the coil for an inductance drop of 10% at the protection limit
Click to expand...

Hi Bruno,

I think that is acceptable practice with distributed gap cores but not with ferrites, unless you design for 100 degrees C or higher. If you look at BH curves for typical power ferrites (N27 etc.) then you can spot that saturation flux density is tipically 100mT lower at 100 degrees than at 25 degrees C. I usually design for 0.25T to 0.3T max to be on the safe side. I can understand economics of mass production, but then you must also state max operating ambient temperature.

Best regards,

Jaka Racman
 
#187
Jaka Racman said:
I think that is acceptable practice with distributed gap cores but not with ferrites, unless you design for 100 degrees C or higher. If you look at BH curves for typical power ferrites (N27 etc.) then you can spot that saturation flux density is tipically 100mT lower at 100 degrees than at 25 degrees C. I usually design for 0.25T to 0.3T max to be on the safe side. I can understand economics of mass production, but then you must also state max operating ambient temperature.
Click to expand...

We must be arguing alongside eachother then. I design for <0.3T as well (lower if high temperature operation is expected).
 
#189
Jaka Racman said:
no problem then, I just wanted to point out that 10% inductance drop is not very well defined quantity with gapped ferrite cores because of steep saturation and temperature dependance.
Click to expand...

Yes, I'm aware of that. Actually the 10% thing is something I use when spec'ing to coil manufs, whereas flux density is what I use when doing the design myself. Coil manufs are notoriously irritable when you go tell them how to design a coil. When you tell them "look I really need a foil winding because of end-to-end capacitance" you still get replies back saying "we will make lits (sic) wire sample OK? Lits is better for HF".

Rrrrrrr....
 
#190
I knocked up a AD180 tonight. Can't say anything as to the sound, haven't gotten it optimized in the biamped rig yet, but it'll be driving bass. I already tweaked it, with non-magnetic hardware throughout, but need to bypass those input caps and maybe the output cap off the opamp besides..... Am I correct in that with the AD opamps, the DC offset off the opamp is very low, and so the cap isn't needed?
 
#191
Yes with the AD op amp, you can get rid of the coupling caps. Provided the audio input signal is DC free of course.

There are coupling caps only at the output of the op amps, not at the input. (in previous versions of the ucd400 the opposite was true). Either way, there are only 2 coupling caps. If you end up shorting 4 caps that's 2 too many 😀
 
#193
Stevenacnj said:
Theres no way we can move this cap off the module and install it like inside the amp and across the + & - speaker binding posts?
Click to expand...
In principle, it could, but you'd also have to move the (differential) feedback takeoff points to the binding posts. Otherwise, the inductance between the module and the filter cap would allow an immense HF to return to the modulator, causing the switching frequency to become anything but the correct value.

In short: I don't recommend it (mildly put).
 
#194
Specifically for Bruno or Jan-Peter:
I got a UCD180AD running, still haven't been able to get much grip on it's sound yet and I'm wanting to add a trafo input. What this does is:
Allows single ended and balanced to be used without shorting pin 1 to pin 3 of the XLR when I want to use RCA. It also ensures that there's no offset going to the amp in case of upstream component failure or user error, so the caps can be bypassed without worrying about the amp being idiotproof. Not to mention the normal positives of using an input trafo.

Here's the rub: The trafo I want to use does not have a center tapped secondary. Shall I use a resistor network to create a virtual ground? Something like 500k from + and - each to ground? Or will the amp function without any true ground connection here so long as + and - are differential?
 
#195
Bruno Putzeys said:

In principle, it could, but you'd also have to move the (differential) feedback takeoff points to the binding posts. Otherwise, the inductance between the module and the filter cap would allow an immense HF to return to the modulator, causing the switching frequency to become anything but the correct value.

In short: I don't recommend it (mildly put).
Click to expand...

So Bruno is this possible to do and if so how would one go about doing it: "you'd also have to move the (differential) feedback takeoff points to the binding posts"

Thanks
 
#196
badman said:
and I'm wanting to add a trafo input.
Click to expand...
"Trafo"? U dutch? Or has Philips lingo spilled over to the English language? 😀
badman said:
The trafo I want to use does not have a center tapped secondary. Shall I use a resistor network to create a virtual ground? Something like 500k from + and - each to ground? Or will the amp function without any true ground connection here so long as + and - are differential?
Click to expand...
The inputs already have 100k from each input to GND so don't bother adding 500k. You should try it but there's a chance that you get some hum if the input side is floating too. A transformer can be relied upon to remove common-mode only if the secondary is indeed grounded. Otherwise. the primary-to-secondary capacitance will neatly couple the common mode into the amp's inputs and you could just as well not have a transformer then. The UcD inputs shouldn't be considered as a truly floating differential input because their common mode range is limited by the op amps and CMRR is not infinite either. So, you may well get better results with the transformer secondary and UcD input connected in a single-ended fashion (ie. one side grounded)
 
#197
Stevenacnj said:
So Bruno is this possible to do and if so how would one go about doing it: "you'd also have to move the (differential) feedback takeoff points to the binding posts"
Click to expand...

It's possible in principle but I've never done it and for the life of it I can't see why you'd want to do such a manoeuvre. It's difficult to imagine what, if any, performance parameter would actually benefit from this.
 
#198
badman said:
Specifically for Bruno or Jan-Peter:
I got a UCD180AD running, still haven't been able to get much grip on it's sound yet and I'm wanting to add a trafo input. What this does is:
Allows single ended and balanced to be used without shorting pin 1 to pin 3 of the XLR when I want to use RCA. It also ensures that there's no offset going to the amp in case of upstream component failure or user error, so the caps can be bypassed without worrying about the amp being idiotproof. Not to mention the normal positives of using an input trafo.

Here's the rub: The trafo I want to use does not have a center tapped secondary. Shall I use a resistor network to create a virtual ground? Something like 500k from + and - each to ground? Or will the amp function without any true ground connection here so long as + and - are differential?
Click to expand...

Badman,
You don’t need Bruno or Jan-peter to answer this one. Each of the inputs already has a resistor to ground of 100k so nothing further needs to be done. However, the specific transformer you are using might like to be terminated in a lower value. This needs to be checked.
Roger
 
Status
Not open for further replies.
Top Bottom