New Amplifier - ULD Extreme

[Carl_Huff]

The ULD Extreme is a new amplifer design inspired by the ULD Amplifier that appeared recently in Silicon Chip magazine. It was spawned from the 'Silicon Chip 200Watt LD amplifier' thread. The ULD Extreme will be a group effort. Assuming that all goes well, this effort will result in built prototypes that will be thoroughly tested and then a group buy for all that are interested.

[jp_howard]

Carl, I'm keen to help with this in any way I can. Since I'm very new, I won't have anything useful to add I expect in terms of design, but if I can help sourcing parts, drawing PCB designs, or testing, please let me know.

The ULD as shown in Silicon Chip's article is a good start, with many people stating that they have built the project and had good results, with the help of some minor mods. Here is the thread that discusses this project and various mods and issues. Here's the main issues I'm aware of:[list=1][*]One area which clearly needs help is bias, which can be improved with a trimpot, or more NJL's, or an extra diode[*]Another area with scope for improvement is the power supply, for which a JLH 'ripple eater' may be useful (amongst other things)[*]An extra diode in the current mirror, along with higher Hfe, good matching in the transistors feeding it, and a trimpot to help match the sides of the current mirror, is another area of improvement identified (although most can be achieved with the existing PCB)[*]Finally, since this is based on Douglas Self's "blameless amp", some areas of distortion mentioned in Self's article that aren't implemented in the SC design need to be addressed[/list=1]I'm sure there's other issues as well, but perhaps this attempt to start the ball rolling by summarising the existing work is a useful starting point for those just coming to this project.

[pheonix358]

I would like to add a wishlist. If it is possible could versions be created with various supply voltages for speakers varying from 2 through to 8 Ohms.

As a loudspeaker designer I have always wanted to design an amp / speaker (active speaker)combination where the amp design can take into account a particular reactive load rather than designing speakers that most amps will run and speakers that are easy on amps. Too many compromises.

Terry

[mickyh]

Originally posted by Carl_Huff
The ULD Extreme is a new amplifer design inspired by the ULD Amplifier that appeared recently in Silicon Chip magazine. It was spawned from the 'Silicon Chip 200Watt LD amplifier' thread. The ULD Extreme will be a group effort. Assuming that all goes well, this effort will result in built prototypes that will be thoroughly tested and then a group buy for all that are interested.

Carl I am the starter of the original SC 200W amplifier thread , quite surprised by all the interest that's been shown in this design.
I am most certainly interested in joining this group effort to build the ULD Extreme
Although i am quite happy with the performance of this Amp so far, i was
considering adding the 2 Extra output Transistors to my existing amps so as to try & remove the "Pot Mod", was also going to try a few of the other members suggestions, but maybe with a new set of Kits from Jaycar as i don't really want
to start modding my Amp that is running so well & sounds great.
F.Y.I. Now running with 12mv across the 0.1 emitters , very stable , runs cool although i am using rather large Heatsinks
So count me in for any help both practical & financial for this project, if we could get some boards fabbed from your new layout i would be quite prepared to obtain the parts reqd & build a pair for test/evaluation.
Could you post the schematic for your new board ?
Did you build these Amps ??
As i am due to retire soon , been looking for a project to keep me amused !!!

Regards Mick Hawtin.

[Carl_Huff]

Please continue posting. Something of consequence will be posted in the next few days. Please be patient.

[jp_howard]

Another proposal for the wish-list - Hugh Dean pointed out that there's very little roll-off at really high frequencies (e.g. 150kHz), which may cause more interference from other sources. He suggested adding a filter, IIRC.

[AKSA]

Mods for Zero Distortion approach:

1. 1N4148 to feed joined bases on current mirror.
2. 2SC1845 for Q8, the EF driving the VAS. Delete the 22K Collector resistor on this device.
3. Reduce Q8 emitter resistor from 2K2 to 470R.
4. Replace Q9 with 2SC3423.
5. Carefully trim lag comp very likely to around 82pF, using a quality silver mica. Too much, slow and leaden, too little, bright and tizzy. Adjust to taste on a variety of genres.
6. Replace the four diode bias string with a BD139 configured as a conventional Vbe multiplier, mounted atop one output device. This gives the option of using other outputs, such as C5200/A1943, which are far cheaper and just as robust and fast.
7. Delete the two 100R resistors between the driver emitters, replace with one 150R paralleled with a 100nF foil polyprop. There is no connection to output rail with either of the two drivers or this 150R resistor.
8. Replace all output device emitter resistors with 0.22R.
9. Match Vbe at 100mA for each output pair within 1mV at 25C.
10. Match also the Hfe of the outputs without 3% at 100mA.
11. As Jeremy suggested, add a 1K/1nF RC filter to the input, to pull back the very high frequency response of this amp, which makes it susceptible to EMI.


No other changes...... that should do it, stable bias, very clear sound, but a tad mechanical sounding. A very nice SS amp with great measurements..... (at 1KHz, compliments to Arthur!).

Cheers,

Hugh

[AKSA]

See above post, OOPS!!

[Carl_Huff]

First question for the group ...

Is there a need for an extra pair of output devices such as what I showed on the PCB that I offered on the 'Silicon Chip 200Watt LD amplifier' thread?

[AKSA]

Carl,

I reckon that'd be a great idea, making six outputs, and the potential for 170W into 8R, and 320W into 4R.

You could build it down to two pairs if you wish; the pcb would simply be a little wider, around 9" instead of 6".

However, since five internal device diodes passing 2mA are required for proper biasing (read: matched tempco so as not to undercompensate thermally), I'd still go for the Vbe multiplier since the VAS current is around 14mA, and it would need to be split between a resistor and the diodes, always a bit risky.....

Cheers,

Hugh