T-Net
I confirmed this with DNM as the datasheet is confusing, the T-net cap includes both 4 pole technology & slit foil technology.
If there ever was a cap to deliver optimum performance, this cap is the one. However I have not heard them yet, mine arrive later this week. I agree with Chris, you would be better off not mixing other caps with these, but it is easy to try, why dont you parallel other caps and report how this sounds?
Regards,
Dean
I confirmed this with DNM as the datasheet is confusing, the T-net cap includes both 4 pole technology & slit foil technology.
If there ever was a cap to deliver optimum performance, this cap is the one. However I have not heard them yet, mine arrive later this week. I agree with Chris, you would be better off not mixing other caps with these, but it is easy to try, why dont you parallel other caps and report how this sounds?
Regards,
Dean
Thank you guys for clarification!
Yes, i also think that T-nets are "too good" for woofer use only. I will take them out from woofer amp and use them elsewhere.
Agree also that T-nets are exellent caps at least with UCD. Have not tried them yet with other amps but i will.
Yes, i also think that T-nets are "too good" for woofer use only. I will take them out from woofer amp and use them elsewhere.
Agree also that T-nets are exellent caps at least with UCD. Have not tried them yet with other amps but i will.
Tantalum caps
Hi All,
Just a reminder about using tantalum caps. If they are used in an enclosed area (pretty warm or hot), their voltage rating can go down 20% or so. Probably a good idea to go overvoltage if you can when using these.
Regards,
Greg
Hi All,
Just a reminder about using tantalum caps. If they are used in an enclosed area (pretty warm or hot), their voltage rating can go down 20% or so. Probably a good idea to go overvoltage if you can when using these.
Regards,
Greg
tantalum caps
Hi Greg,
These caps to replace the ceramics, are just in the Plus and Minus lines of the 12V rail. That should help.
Ray
Hi Greg,
These caps to replace the ceramics, are just in the Plus and Minus lines of the 12V rail. That should help.
Ray
correct transformer sizes
Hi All,
Let me say first of all I am not again going through all of the forums and take another 3 days to try and find out the correct answer, if there is one.
First off, for both the UCD180 and UCD400, is there a correct size for a mono setup for each versus using 1 transformer with 2 bridges caps ETC.? I have read somewhere that to big a transformer will slow down the current draw, versus a slightly smaller one will let current flow faster because of the metal in the transformer. It's a comprimise. Ehnce my question.
Secondly, The Avels from Part Express appear to be aok. Any experiences good or bad.
3. One last question, not transformer related, does anyone know if the newest version 6.1 AD180 has that DCR on it like the AD400?
Ray
Hi All,
Let me say first of all I am not again going through all of the forums and take another 3 days to try and find out the correct answer, if there is one.
First off, for both the UCD180 and UCD400, is there a correct size for a mono setup for each versus using 1 transformer with 2 bridges caps ETC.? I have read somewhere that to big a transformer will slow down the current draw, versus a slightly smaller one will let current flow faster because of the metal in the transformer. It's a comprimise. Ehnce my question.
Secondly, The Avels from Part Express appear to be aok. Any experiences good or bad.
3. One last question, not transformer related, does anyone know if the newest version 6.1 AD180 has that DCR on it like the AD400?
Ray
maxlorenz said:
Just mod. it.
An externally hosted image should be here but it was not working when we last tested it.
First listening test shows no change comparing it directly to standard amp. Has to burn in maybe? I mean the tantaliums.
Dear Bert: 🙂
I am extremelly grateful!
Do you kept the stock caps?
Can you (or anyone) figure out the whole input section. It is not apparent to me.
That has never been posted, AFAIK.
Maybe it's worth further mods...
See, I run my UCD400ST DC coupled, with direct soldering of signal wires to (former) coupling cap's holes. I still have HF distortion no matter what I try. Maybe I fried something 😀
Best of lucks.
Mauricio
I am extremelly grateful!
Do you kept the stock caps?
Can you (or anyone) figure out the whole input section. It is not apparent to me.
That has never been posted, AFAIK.
Maybe it's worth further mods...
See, I run my UCD400ST DC coupled, with direct soldering of signal wires to (former) coupling cap's holes. I still have HF distortion no matter what I try. Maybe I fried something 😀
Best of lucks.
Mauricio
Mauricio,
You mean this? This is only with a NE5532.
http://www.grotel.nl/photo/input.gif
I kept the original cer. caps also.
What do you mean HF distortion? Is the input signal too high? Do you use a preamp or something else? As you can see on the picture I also have no input caps. Do you have distortion in both channels. Is it possible to show/write how you've connected/wired your amp.?
You mean this? This is only with a NE5532.
http://www.grotel.nl/photo/input.gif
I kept the original cer. caps also.
What do you mean HF distortion? Is the input signal too high? Do you use a preamp or something else? As you can see on the picture I also have no input caps. Do you have distortion in both channels. Is it possible to show/write how you've connected/wired your amp.?
ive used several avel's without a problem. they seem to work just fine. they are a great price too. one recommendation here!
I've bought six Avel's so far and have never had any problems. Although it is not obvious from their webpage, you can order directly from them, at competitive prices, by contacting their sales dept. This gives you access to their full line-up of values, PE only carries certain ranges.
Paul
Paul
Re: correct transformer sizes
Ray,
To be conservative, try to choose the VA of the transformer so the peak currents in the supply are less than the peak current the transformer is rated at into a resistive load.
If its one tranny, two windings, one for each rail, its okay to assume you have the full VA for each winding as both will not be pulling big at the same time.
Also, keep in mind that with a class D amp, more voltage on the rails is good. It reduces ripple, and the current the tranny is required to supply. The class D draws energy, not current, so more volts means more 1/2 C V^2 energy.
I've found that 400VA is plenty for one UcD400 in real life into 6 ohms, and likely fine into 4 ohms. I'd even go so far as to say if you can only muster 400VA, its likely fine for two channels into 4 ohms. It may not be ideal though. Key point is to design the system so the peak currents demanded are less than the amp rating of the core (usually both in parallel) times 1.41.
You really need to simulate to get there and I've attatched the one I've used modeling the transformer based on its posted (and very detailed) specs and experienced guestimates.
If your looking to size a xfrmer, here is a spice sim I did that models the demands of an ideal class D as a load.
Save it, open it, rename it, and polk around with the output voltage, load resistance, VAC input, winding resistance, and output cap to your liking.
Can't help much with it as I'm swamped.
The xfrmer modeled is this one ( assuming one winding) and is for only half the power supply, but the other half is the same. (i.e. dual diode bridge config) This was modeled for a monoblock, but I think you will find it will be fine for two channels, if that's a compromise you wanna make.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7540251656&rd=1&sspagename=STRK:MEWA:IT&rd=1
please note, that some will say this is to aggressive to use a 46V tranny, but sim's show otherwise, and keep in mind, the UcD's shut down if the voltage rises, so overvoltage is not an issue. That is to say, if you get say 130VAC plus some day, your UcD's just will shut down, and the fets are good to 150V each in this state. Point being, if it gets that bad, you've got problems no matter what, but your amp will work when the voltage drops.
Please note the boosting affect changes with output excursion, and can be modeled. I figured it out a while ago, I think its at 50% output, but it may be 75%, don't have time to dig at it. Sim is quick, so you can see it.
Note the boosting adds 2nd harmonic to the rail!
Look at V+, Vout(what your speakers get) and I(R1), which is most important when sizing the VA of the tranny to assure it doesn't saturate (although that's my biased opinion.)
At 20Hz Vout, I assumed this worst case. Records will be much worse with rumble though.
Regards,
Mike
p.s. The attatched simulation file can be opened with linear tech's free Spice (SitcherCAD III). www. linear.com
Its almost instantaniously fast since the class D power consumptin has been idealized.
ray bronk said:
Ray,
To be conservative, try to choose the VA of the transformer so the peak currents in the supply are less than the peak current the transformer is rated at into a resistive load.
If its one tranny, two windings, one for each rail, its okay to assume you have the full VA for each winding as both will not be pulling big at the same time.
Also, keep in mind that with a class D amp, more voltage on the rails is good. It reduces ripple, and the current the tranny is required to supply. The class D draws energy, not current, so more volts means more 1/2 C V^2 energy.
I've found that 400VA is plenty for one UcD400 in real life into 6 ohms, and likely fine into 4 ohms. I'd even go so far as to say if you can only muster 400VA, its likely fine for two channels into 4 ohms. It may not be ideal though. Key point is to design the system so the peak currents demanded are less than the amp rating of the core (usually both in parallel) times 1.41.
You really need to simulate to get there and I've attatched the one I've used modeling the transformer based on its posted (and very detailed) specs and experienced guestimates.
If your looking to size a xfrmer, here is a spice sim I did that models the demands of an ideal class D as a load.
Save it, open it, rename it, and polk around with the output voltage, load resistance, VAC input, winding resistance, and output cap to your liking.
Can't help much with it as I'm swamped.
The xfrmer modeled is this one ( assuming one winding) and is for only half the power supply, but the other half is the same. (i.e. dual diode bridge config) This was modeled for a monoblock, but I think you will find it will be fine for two channels, if that's a compromise you wanna make.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7540251656&rd=1&sspagename=STRK:MEWA:IT&rd=1
please note, that some will say this is to aggressive to use a 46V tranny, but sim's show otherwise, and keep in mind, the UcD's shut down if the voltage rises, so overvoltage is not an issue. That is to say, if you get say 130VAC plus some day, your UcD's just will shut down, and the fets are good to 150V each in this state. Point being, if it gets that bad, you've got problems no matter what, but your amp will work when the voltage drops.
Please note the boosting affect changes with output excursion, and can be modeled. I figured it out a while ago, I think its at 50% output, but it may be 75%, don't have time to dig at it. Sim is quick, so you can see it.
Note the boosting adds 2nd harmonic to the rail!
Look at V+, Vout(what your speakers get) and I(R1), which is most important when sizing the VA of the tranny to assure it doesn't saturate (although that's my biased opinion.)
At 20Hz Vout, I assumed this worst case. Records will be much worse with rumble though.
Regards,
Mike
p.s. The attatched simulation file can be opened with linear tech's free Spice (SitcherCAD III). www. linear.com
Its almost instantaniously fast since the class D power consumptin has been idealized.
maxlorenz said:
By bypassing the input section you are running directly in the UcD modulator, with has input impedances of under 10k (don't remember the exact values). HF distortion increases with decreasing of the input impedance, also you might be passing on some MHZ range garbage directly to the UcD modulator.
The UcD input section has 100k impedance and also some 100pf caps to shunt high frequency content, so it would be advisable to use the original input pins.
maxlorenz said:
M,
I know you love those filterless Dac's, but you can't be doing that into a UcD IMO!
Mike
maxlorenz said:
M,
I know you love those filterless Dac's, but you can't be doing that into a UcD IMO!
Mike
correct size of Transformer
Hi Portlandmike,
On the UCD400, the cutoff or shutdown voltage is 68V. First of all, "If its one tranny, two windings, one for each rail, its okay to assume you have the full VA for each winding as both will not be pulling big at the same time." If that be the case, why use 2 transformers 1 for each side. Isn't there some interaction between windings, since the ground wire for both secondary windings has to be shared?
Secondly, 45x1.414 = 63.63, just over the voltage of some of the caps on the board. Now somewhere I do remember reading a discussion about attempting to figure out maximum voltage needed to take in to account the bridge caps and actual load. 10% after that is about 69V. haven't measured the actual AC voltage, but if it is more than 120V, wouldn't that be a problem? Just concerned that I'm running to close to the rails.
Ray
Hi Portlandmike,
On the UCD400, the cutoff or shutdown voltage is 68V. First of all, "If its one tranny, two windings, one for each rail, its okay to assume you have the full VA for each winding as both will not be pulling big at the same time." If that be the case, why use 2 transformers 1 for each side. Isn't there some interaction between windings, since the ground wire for both secondary windings has to be shared?
Secondly, 45x1.414 = 63.63, just over the voltage of some of the caps on the board. Now somewhere I do remember reading a discussion about attempting to figure out maximum voltage needed to take in to account the bridge caps and actual load. 10% after that is about 69V. haven't measured the actual AC voltage, but if it is more than 120V, wouldn't that be a problem? Just concerned that I'm running to close to the rails.
Ray
Hi Mighty Mike and Lucpes 😀
Here I am, with 4 or 5 unfinished projects...
I know you warned me, yet my other system (16*TDA1543 to UCD180) is quiet ( with your advice 😉 ) and delta sigma DACs are not welcome either by the disfunctional amp...
I will soon return to my, now completelly super-regulated, superDAC 😎 ( I made a mistake with first ALWSR: I forgot one cap!
)
Thanks for your interest. Your diagram is for UCD180. My trouble is with UCD400ST, wich has a cheap elco as signal coupling cap. This cap is the only one that I removed and signal is fed directly to its hole, bypassing only the input connector...and some cms of via...and 2 soldering points 😉 Did I make something wrong?
If you look at Bgt's pick you could follow trace from input connector to elco. There's nothing in between, wright?
The SMD input section has been respected.
I will try pi filter at main PS (as per Rha61 advice) , because 47uF/47R snubber almost eliminated distortion but at the cost of diminished transparency and speed. Other snubber values led to distinct but unworthy changes. It's hard without scope and knowledge. Perhaps will try shielded input wires.
Best regards.
Mauricio
PS: RC (10nF/560R) to the secondaries of my TVC didn't help either
PS nº2: I received a mail from National about "Half-Bridge Power Converter Design Using the LM5035", wich seems similar to what you (Mike) were talking about recently (?)
How do you do?
Here I am, with 4 or 5 unfinished projects...
I know you warned me, yet my other system (16*TDA1543 to UCD180) is quiet ( with your advice 😉 ) and delta sigma DACs are not welcome either by the disfunctional amp...
I will soon return to my, now completelly super-regulated, superDAC 😎 ( I made a mistake with first ALWSR: I forgot one cap!
)Thanks for your interest. Your diagram is for UCD180. My trouble is with UCD400ST, wich has a cheap elco as signal coupling cap. This cap is the only one that I removed and signal is fed directly to its hole, bypassing only the input connector...and some cms of via...and 2 soldering points 😉 Did I make something wrong?
If you look at Bgt's pick you could follow trace from input connector to elco. There's nothing in between, wright?
The SMD input section has been respected.
I will try pi filter at main PS (as per Rha61 advice) , because 47uF/47R snubber almost eliminated distortion but at the cost of diminished transparency and speed. Other snubber values led to distinct but unworthy changes. It's hard without scope and knowledge. Perhaps will try shielded input wires.
Best regards.
Mauricio
PS: RC (10nF/560R) to the secondaries of my TVC didn't help either
PS nº2: I received a mail from National about "Half-Bridge Power Converter Design Using the LM5035", wich seems similar to what you (Mike) were talking about recently (?)
Hey Mauricio,
If you are using the standard hypex, UCD400, you should have either a 2134 or an AD8620. If 2134, check for voltage offset. To much can cause distortion. Also, maybe coming off the DAC, maybe to much signal at imput. Just some lame suggestions.
Ray
If you are using the standard hypex, UCD400, you should have either a 2134 or an AD8620. If 2134, check for voltage offset. To much can cause distortion. Also, maybe coming off the DAC, maybe to much signal at imput. Just some lame suggestions.
Ray
maxlorenz said:
It's more or less the same input section as I posted for the 400ST. If you look closely at the schematic, the opamp is BEFORE the coupling caps.
Hello all,
I've been reading, reading, searching, reading and then reading some more - followed by some reading and a little reading and then some.....
.....this has to be one of the most popular threads ever, judging from the number of posts. And I have now sort of given up on ever getting to magic overview of it all....
(Not to mention one or two other threads that is considered madatory reading on the UCDs as well!)
So my Q is: Has anyone sort of digested/compiled an outline plan of things to mod/what works within all this! Pointeres to posts, webpages etc. would be highly appreciated in this matter. Who will/can serve it on a silver plate for us lazy guys...🙄
PS: My urge is to get the most out of the 180 module for dedicated mid use in active 3-way system (150 Hz to 2.5 KHz)
I've been reading, reading, searching, reading and then reading some more - followed by some reading and a little reading and then some.....
.....this has to be one of the most popular threads ever, judging from the number of posts. And I have now sort of given up on ever getting to magic overview of it all....
(Not to mention one or two other threads that is considered madatory reading on the UCDs as well!)So my Q is: Has anyone sort of digested/compiled an outline plan of things to mod/what works within all this! Pointeres to posts, webpages etc. would be highly appreciated in this matter. Who will/can serve it on a silver plate for us lazy guys...🙄
PS: My urge is to get the most out of the 180 module for dedicated mid use in active 3-way system (150 Hz to 2.5 KHz)
- Status
- Not open for further replies.