UCD180 questions

ghemink · 21st July 2004, 01:13 PM

Hello Jan-Peter, Bruno,

As this subject maybe of interest to more readers, just a few questions here on the Hypex UCD180 modules.

I received 4 UCD180 modules today. Apparently I`m one of the first customers since I had serial number 0004 to 0007 :-)
I have not yet hooked them up to a power supply, so I have not done any listening tests (will have to wait until weekend or even later).

A few questions concerning the input of the module. The module has a symmetric input using a NE5532 dual opamp. It seems to be used as an instrumental amplifier like input with the UCD itself acting as the 3rd differential amp. I can not see all traces on the PCB but it seems to be connected like an instrumental amplifier with a gain of about 4.5 or so. The outputs of the two opamps go to the inputs of the UCD amp itself. This is done by two coupling capacitors that are 22uF electrolitic capacitors. Assuming that the input impedance of the UCD amp itself is something like 1.8k for the inverting input (as stated in the classd.pdf file), then that gives a high pass filter at 4Hz. That should be fine for all apllications. However, when I`m going to use this amp for a tweeter or a midrange, I sure would want to replace the electrolitic caps with a different type of cap and of course with a lower value. Are those caps really required? What happens if they are replaced with a piece of wire (0 Ohm resistor). Would that give problems when the input voltage is not purely symmetrical? I assume that when the input signal is pure symmetrical, it sould not hurt to remove the caps, would it?

Best regards

Gertjan

phase_accurate · 22nd July 2004, 06:41 AM

Quote:

However, when I`m going to use this amp for a tweeter or a midrange, I sure would want to replace the electrolitic caps with a different type of cap and of course with a lower value.

Keep in mind that this will give another pole in the highpass transfer function if you take a smaller one.
So I would either leave them as they are (for all channels) in order to not have to care about this added pole or deliberately dimension them such that they are a part of the wanted x-over function.
On the ones that I will get, I will probably add a parallel polyprop cap.
But I would never replace them by a piece of wire.

To be sure if it is save to remove them you may ask Jan Peter by e-mail and post his answer here.

Regards

Charles

ghemink · 22nd July 2004, 07:07 AM

Hi Charles,

Thanks for your reply. I know I will add another pole that may influence the x-over behavior when I change those caps. However, I can take a safe margin, for example make a high-pass for the tweeter at 500Hz, so the cap can be reduced in size with a factor 100. This also makes the whole thing a bit more robust, less likely to blow up the tweeter when something weird happens at the input of the amp. For the midrange, I was thinking of making the pole at 50Hz (crossover in the order of 120-150Hz). However, there is something tricky here. In the file classd.pdf, it is stated that the input impedance of the + and - input of the UCD amp itself is different. This means that the poles for + and - input will be at a different point. This would give a different behaviour depending on whether your input signal is symmetrical or unipolar. This may not be so nice. However, could be solved (if needed, will not do that at this stage) likely by deliberately lowering the input impedance of that UCD input that has the highest inout impedance (to make them equal).

However, in my setup, I plan to use the modules driven symmetrically by my DEQX PDC. The DEQX PDC measures the drivers frequency and phase response (including that of the amp)and corrects for it and applies phase linear crossovers. In this setup, another pole that is relatively close to the crossover (say 2-3 octaves a way) does not hurt at all. The PDC would compensate for it.

Once you have the modules, let me know your expereinces with them, it may take sometime before I`m able to do listening tests. I`ll try in the weekend. Listening tests will first be with a passive speaker system (B&W CDM9NT) comparing them with my Accuphase E407 integrated amp.

I think Bruno and Jan-Peter will react (they reacted to my previous posts on this site). If not, I`ll contact them directly. I guess these caps and possibly the NE5532 would be welcome objects for tweaking. But first I`ll keep everything intact before running to my soldering iron :-)

Best regards

Gertjan

UrSv · 22nd July 2004, 01:37 PM

Quote:

Originally posted by ghemink
-- snip --
I received 4 UCD180 modules today. Apparently I`m one of the first customers since I had serial number 0004 to 0007 :-)
-- snip --

I got serial numbers 0008 and 0009 so I'm right behind you...

phase_accurate · 22nd July 2004, 02:10 PM

S**t

I got 12 to 17.

Regards

Charles

UrSv · 22nd July 2004, 02:14 PM

Slow, heh? So...who got 0010 and 0011?

Jan-Peter · 22nd July 2004, 04:59 PM

Guys,

I am very pleased to mine serial numbers back

We started also realy with the number 1 and not with 1001, as some other manufactures will do.

Please don't remove the cap, because you will have the risk of powersupply pumping effect. When you remove the cap please make use of an overvoltage protection.

A good idee for a tweak would be a Black Gate 22uF and an OP2134.

For an other customer we put an AD8620 (perhaps the next serial numers?) , but they are quit costly......

Please don't change anything before you have listen to the UcD180!

If it is possible please use the balanced input. When you don't have a balanced output, use a symmetrical cable and connect the -IN at ground at the end of the cable not by the UcD180 module!

Reagrds,

Jan-Peter

www.hypex.nl

phase_accurate · 22nd July 2004, 06:37 PM

When were at it:

What are the input impedances of the two inputs ? I am asking because I may probably use a 1st order passive RC crossover on one of the two channels.

Regards

Charles

Jan-Peter · 22nd July 2004, 07:08 PM

Charles,

Please take a look on the attachment. The inputimpedance of the UcD modulator is 1.8K.

Regards,

Jan-Peter

www.hypex.nl

phase_accurate · 22nd July 2004, 07:17 PM

It was not the input impedance of the modulator per se I was interested in but the input buffer. But from the diagram I can see that it is around 100k from each "leg" to ground (at least at LF).

Thanks

Charles

21st July 2004, 01:13 PM	#1
ghemink diyAudio Member Join Date: Jun 2004 Location: Japan	UCD180 questions Hello Jan-Peter, Bruno, As this subject maybe of interest to more readers, just a few questions here on the Hypex UCD180 modules. I received 4 UCD180 modules today. Apparently I`m one of the first customers since I had serial number 0004 to 0007 :-) I have not yet hooked them up to a power supply, so I have not done any listening tests (will have to wait until weekend or even later). A few questions concerning the input of the module. The module has a symmetric input using a NE5532 dual opamp. It seems to be used as an instrumental amplifier like input with the UCD itself acting as the 3rd differential amp. I can not see all traces on the PCB but it seems to be connected like an instrumental amplifier with a gain of about 4.5 or so. The outputs of the two opamps go to the inputs of the UCD amp itself. This is done by two coupling capacitors that are 22uF electrolitic capacitors. Assuming that the input impedance of the UCD amp itself is something like 1.8k for the inverting input (as stated in the classd.pdf file), then that gives a high pass filter at 4Hz. That should be fine for all apllications. However, when I`m going to use this amp for a tweeter or a midrange, I sure would want to replace the electrolitic caps with a different type of cap and of course with a lower value. Are those caps really required? What happens if they are replaced with a piece of wire (0 Ohm resistor). Would that give problems when the input voltage is not purely symmetrical? I assume that when the input signal is pure symmetrical, it sould not hurt to remove the caps, would it? Best regards Gertjan

22nd July 2004, 02:14 PM	#6
UrSv diyAudio Member Join Date: Jan 2002 Location: Sweden/Switzerland	Slow, heh? So...who got 0010 and 0011? __________________ UrSv Those who say it can't be done should not stop those who are doing it.

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

22nd July 2004, 07:07 AM	#3
ghemink diyAudio Member Join Date: Jun 2004 Location: Japan	Hi Charles, Thanks for your reply. I know I will add another pole that may influence the x-over behavior when I change those caps. However, I can take a safe margin, for example make a high-pass for the tweeter at 500Hz, so the cap can be reduced in size with a factor 100. This also makes the whole thing a bit more robust, less likely to blow up the tweeter when something weird happens at the input of the amp. For the midrange, I was thinking of making the pole at 50Hz (crossover in the order of 120-150Hz). However, there is something tricky here. In the file classd.pdf, it is stated that the input impedance of the + and - input of the UCD amp itself is different. This means that the poles for + and - input will be at a different point. This would give a different behaviour depending on whether your input signal is symmetrical or unipolar. This may not be so nice. However, could be solved (if needed, will not do that at this stage) likely by deliberately lowering the input impedance of that UCD input that has the highest inout impedance (to make them equal). However, in my setup, I plan to use the modules driven symmetrically by my DEQX PDC. The DEQX PDC measures the drivers frequency and phase response (including that of the amp)and corrects for it and applies phase linear crossovers. In this setup, another pole that is relatively close to the crossover (say 2-3 octaves a way) does not hurt at all. The PDC would compensate for it. Once you have the modules, let me know your expereinces with them, it may take sometime before I`m able to do listening tests. I`ll try in the weekend. Listening tests will first be with a passive speaker system (B&W CDM9NT) comparing them with my Accuphase E407 integrated amp. I think Bruno and Jan-Peter will react (they reacted to my previous posts on this site). If not, I`ll contact them directly. I guess these caps and possibly the NE5532 would be welcome objects for tweaking. But first I`ll keep everything intact before running to my soldering iron :-) Best regards Gertjan

22nd July 2004, 02:10 PM	#5
phase_accurate diyAudio Member Join Date: May 2002 Location: Switzerland	S**t I got 12 to 17. Regards Charles

22nd July 2004, 04:59 PM	#7
Jan-Peter diyAudio Member Join Date: Sep 2003 Location: -	Guys, I am very pleased to mine serial numbers back We started also realy with the number 1 and not with 1001, as some other manufactures will do. Please don't remove the cap, because you will have the risk of powersupply pumping effect. When you remove the cap please make use of an overvoltage protection. A good idee for a tweak would be a Black Gate 22uF and an OP2134. For an other customer we put an AD8620 (perhaps the next serial numers?) , but they are quit costly...... Please don't change anything before you have listen to the UcD180! If it is possible please use the balanced input. When you don't have a balanced output, use a symmetrical cable and connect the -IN at ground at the end of the cable not by the UcD180 module! Reagrds, Jan-Peter www.hypex.nl

22nd July 2004, 06:37 PM	#8
phase_accurate diyAudio Member Join Date: May 2002 Location: Switzerland	When were at it: What are the input impedances of the two inputs ? I am asking because I may probably use a 1st order passive RC crossover on one of the two channels. Regards Charles

22nd July 2004, 07:17 PM	#10
phase_accurate diyAudio Member Join Date: May 2002 Location: Switzerland	It was not the input impedance of the modulator per se I was interested in but the input buffer. But from the diagram I can see that it is around 100k from each "leg" to ground (at least at LF). Thanks Charles

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)