OK- hopefully a photo of the underside of the coil is attached (haven't tried attaching anything before!)
Doesn't reveal very much more - you can see the winding start and finish, and that it's a solid wire
you can see the blobs of epoxy holding the two halves together (and where I have tried and failed to remove the nearest one), so unless I destroy it, we can only guessimate the gap spacing.
So - It's an EP17 core, apparently wound with approx 22 turns of 22 swg wire - it's value is 30uH, and it seems to be rated at 180 Watts.
From what I can see, the UCD180 coil looks very similar - except for the mounting pins, which have been removed completely from one side of the former. It is only attached to the pcb by the pins on the wire termination side, and the clamp.
Doesn't reveal very much more - you can see the winding start and finish, and that it's a solid wire
you can see the blobs of epoxy holding the two halves together (and where I have tried and failed to remove the nearest one), so unless I destroy it, we can only guessimate the gap spacing.
So - It's an EP17 core, apparently wound with approx 22 turns of 22 swg wire - it's value is 30uH, and it seems to be rated at 180 Watts.
From what I can see, the UCD180 coil looks very similar - except for the mounting pins, which have been removed completely from one side of the former. It is only attached to the pcb by the pins on the wire termination side, and the clamp.
I'm considering trying out a small UcD based on the App note from Philips. (It truely is a circuit of beauty 
)
Target will be some 200W/8Ohm which should be possible with a 2x40VAC /2x56VDC supply.
To make it cheap and easy to obtain, I will try to substitude some of the components, with more general parts (components that can be obtained through RS-components).
I base it on SMD components to make it as small as possible, both to save cost on PCB, but also to minimize noise coupling problems generated by the switching currents.
Zilog/RX5 are you using Schottky diodes in the comparator and in the driver?
Maybe using signal diodes like 1N4148 would do!
the transistors in the drivers (PBSS5140T) is used af these are pretty fast and can handle a lot of current (2A).
The FZT790A Zilog is using is not a bad choice.
Think the overall gain is way to small, and also the input impedance is to small. Zilog/RX5 what gain have you settled on? Are you using a buffer stage at the input.
Adding a buffer will unfortunatly raise the component count, as it will also require a dual supply to support it (if an opamp is used). Maybe I could just make a simple buffer with a couple of BJTs.
I thinkI'l go for a toroid output coil, but the PCB could maybe be made to accept an RM core also! The PCB shopuld be dual layer with all components (no vertical PCB for the input) .... must be possible to make it so that the coil is not coupling too much noise into the circuit!
If a very simple current limiter (shot off) could be implemented I think it shoud be included as well!
Main goal is to make a good and cheep amp, which could be a platform for more experiments.
I could make the PCB files available in gerber format, and anyone interested could just order these from e.g. Olimex.
I think it could fit on a quarter of a Euro board, making it less that 10$ a pice.
To save me both money and time, I would need someone to review both PCB layout and circuit.
Anyone interested in participating in such a project?
)Target will be some 200W/8Ohm which should be possible with a 2x40VAC /2x56VDC supply.
To make it cheap and easy to obtain, I will try to substitude some of the components, with more general parts (components that can be obtained through RS-components).
I base it on SMD components to make it as small as possible, both to save cost on PCB, but also to minimize noise coupling problems generated by the switching currents.
Zilog/RX5 are you using Schottky diodes in the comparator and in the driver?
Maybe using signal diodes like 1N4148 would do!
the transistors in the drivers (PBSS5140T) is used af these are pretty fast and can handle a lot of current (2A).
The FZT790A Zilog is using is not a bad choice.
Think the overall gain is way to small, and also the input impedance is to small. Zilog/RX5 what gain have you settled on? Are you using a buffer stage at the input.
Adding a buffer will unfortunatly raise the component count, as it will also require a dual supply to support it (if an opamp is used). Maybe I could just make a simple buffer with a couple of BJTs.
I thinkI'l go for a toroid output coil, but the PCB could maybe be made to accept an RM core also! The PCB shopuld be dual layer with all components (no vertical PCB for the input) .... must be possible to make it so that the coil is not coupling too much noise into the circuit!
If a very simple current limiter (shot off) could be implemented I think it shoud be included as well!
Main goal is to make a good and cheep amp, which could be a platform for more experiments.
I could make the PCB files available in gerber format, and anyone interested could just order these from e.g. Olimex.
I think it could fit on a quarter of a Euro board, making it less that 10$ a pice.
To save me both money and time, I would need someone to review both PCB layout and circuit.
Anyone interested in participating in such a project?
lumanauw said:Hi, Rogs,
Thanks alot
Do you have an inductance-meter? I know it is written 30mH, but could you measure it's actual inductance with inductance-meter?
It's 30uH, not 30mH, on the UCD180 coil label - the Philips one doesn't have a value written on it, but is listed as 28uH on the circuit diagram.
I don't have an inductance meter, but I should be able to measure the inductance using my signal generator at work, and a series resistor.
I would expect the Z of the coil to be somewhere around 17.5R at 400 KHz.
I'll check next week.
good luck rogs said:
The inductor in the photo (the Philips one) measures out at a fraction over 24uH, averaged over several frequencies, from 100KHz to 400KHz. I measured the resonant frequency by fitting a selection of capacitors across the coil, and looking for the peak response on the scope.
Nearest I have to an inductance meter!
Obviously the output inductor values can have a degree of flexibility in actual realisation --- my UCD180 has 30uH printed on the label, the Philips has 28uH on the schematic, and is actually 24uH on the PCB .
But how extreme could the variations be?
Say I decided to use a filter with an Fc of c. 34KHz :
I could choose 47uH and .47uF
Or 4.7uH and 4.7uF
Surely the only difference would be in the Q of the filter, with the latter being 10 times bigger.
But as Bruno points out in his AES paper, the physical resonance of the filter does not figure at all in the closed loop response, and enclosing the filter in the single control loop ensures that any gain boost ptoduced by it is subsequently cancelled by an equal boost in the loop gain.
So is there any thing to gain by making the inductor value bigger (apart from smaller capacitors!!)? ---
It's just that there's a range of Epcos SMD inductors that are fairly easy to get, but they have a reasonably low Isat.
I only need a max current of about 5A , so it would be better to use the 10uH which has an Isat of 6.4A, rather than a 47uH from the same range, which only has an Isat of 3A.
Unless of course I'm missing something altogether ?
But how extreme could the variations be?
Say I decided to use a filter with an Fc of c. 34KHz :
I could choose 47uH and .47uF
Or 4.7uH and 4.7uF
Surely the only difference would be in the Q of the filter, with the latter being 10 times bigger.
But as Bruno points out in his AES paper, the physical resonance of the filter does not figure at all in the closed loop response, and enclosing the filter in the single control loop ensures that any gain boost ptoduced by it is subsequently cancelled by an equal boost in the loop gain.
So is there any thing to gain by making the inductor value bigger (apart from smaller capacitors!!)? ---
It's just that there's a range of Epcos SMD inductors that are fairly easy to get, but they have a reasonably low Isat.
I only need a max current of about 5A , so it would be better to use the 10uH which has an Isat of 6.4A, rather than a 47uH from the same range, which only has an Isat of 3A.
Unless of course I'm missing something altogether ?
I would like to contact Steven who first posted the circuit of the Philips UCD. Unfortunately I am not allowed to contact him because I have not posted enough replies yet - these will come.
The design looks very sound and there is no reason to shoot it down by anyone due to selfishness or jealousy, or by someone who would like to have this thread to sleep in silence. Negative remarks will only be ignored anyhow. Can someone bring me into contact with Steven? An e-maill address will be preferred. Thank you.
The design looks very sound and there is no reason to shoot it down by anyone due to selfishness or jealousy, or by someone who would like to have this thread to sleep in silence. Negative remarks will only be ignored anyhow. Can someone bring me into contact with Steven? An e-maill address will be preferred. Thank you.
Re: Philips UCD
I have tried both the Philips Demonstration PCB, and built my own version of the circuit.
It works OK - up to a point. The non differential input causes a lot of instability, especially with loads of less than 8R. I cannot see how Philips have managed to get the 200W into 4R that they claim - in my experience the PSU rail pumping due to DC offset is just too great.
The early posts in this thread refer to the non differential input problem -- It's something that needs to have careful attention paid to it, IMHO.
If you want to try this circuit, 'go differential'!
stepheno said:
I have tried both the Philips Demonstration PCB, and built my own version of the circuit.
It works OK - up to a point. The non differential input causes a lot of instability, especially with loads of less than 8R. I cannot see how Philips have managed to get the 200W into 4R that they claim - in my experience the PSU rail pumping due to DC offset is just too great.
The early posts in this thread refer to the non differential input problem -- It's something that needs to have careful attention paid to it, IMHO.
If you want to try this circuit, 'go differential'!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.