Unchoked the sound of SI 5066

Panelhead · « **on:** 19 Feb 2005, 03:13 am »

I rebuilt two of the SI 5066 recently. One had the main filter cap upgraded and 4.7 ufd coupling caps added. The other has those two changes and the stock chokes were replaced with 10 uH, 1500 ma chokes made by Miller.
The unit with the new chokes did not get used for a while, there was a problem with my soldering. Listened to the other board with the bigger cap and the 0.33 ufd caps replaced with 4.7 ufd multilayer ceramics. The higher value caps did fill in the low octaves. Since most listening is either classical or NPR, the improvement was music dependent. In some cases more extended low end was not used. But overall it was a big improvement over a stock SI board.
I sorted out the mistake in the other board that has the new chokes and inserted it into my chassis. Now this change is audible with all music and programs. The clarity and detail of the 2024 chip is still there in spades, but the dynamics are even bigger. The difference between soft and loud has gotten a lot wider. The dynamic shadings are more pronounced.
My power demends are minimal, the speakers are full range, 100 dB, 15 ohm. The listening level is mids 80's peak, so 100 mW or so.
I suspect the small stock chokes are doing more than stopping RFI. There may be introducing some dynamic compression.
The chokes used are nothing special, they are 1.00 each locally. Digikey does not list this size, but Newark does. They are about 1.50 each at Newark.
From a listening standpoint, the change in chokes seems to make a bigger improvement in sonics than bumping up the onboard filter cap. Maybe more than changing the filter caps.
Wish I could figure out how to post pics.

George

Feisal K · « **Reply #1 on:** 19 Feb 2005, 03:45 am »

George, open an account at http://www.photobucket.com/ you can upload pics there and then copy the whole [ IMG ] ... [/ IMG ] tag and paste it here.

dwk · « **Reply #2 on:** 19 Feb 2005, 06:34 am »

I assume these are air-core chokes?

I got into digital amps with the Panny units, and am still baffled by the fact that folks spent time on wire and speaker jacks while the output inductors were ignored - good air-core inductors (perhaps with some shielding improvements to mitigate any extra rfi) would seem to be a no-brainer, aside from the mechanical/space problems.

Panelhead · « **Reply #3 on:** 19 Feb 2005, 11:23 am »

Quote from: dwk

I assume these are air-core chokes?

I got into digital amps with the Panny units, and am still baffled by the fact that folks spent time on wire and speaker jacks while the output inductors were ignored - good air-core inductors (perhaps with some shielding improvements to mitigate any extra rfi) would seem to be a no-brainer, aside from the mechanical/space problems.

No, these J.W.Miller chokes are ferrite cored also. The current rating is lower than the factory units, but they are about 4 -5 times larger. The wire used to wrap the core is much larger also. Must be the difference a between a RF choke and a power choke.
The size has required mounting under the circuit board. But these are small compared to the nice 15 uH Solo chokes I picked up for my next project using the SI board. I suspect the air cored Solo inductors will prove to be even better. They are made for speaker crossovers. The 16 ga size I have is discontinued now, and it is much heavier than what is needed.
One downside from my playing with these. I think maybe the input cap will get reduced in size to 2.2 ufd like the Tripath data sheet shows. There is a pop on turn on now that came with the input cap change. May be the charging up of the cap as the bias voltage is applied. A smaller value cap will charge up quicker. Lower sensitivity speakers than mine would be less effected.

George

BillyM · « **Reply #4 on:** 19 Feb 2005, 06:52 pm »

What version of the 5066 do you have? Also, would you be so kind as to toss up a part-number and source for your upgrade components? I can enjoy music, I can solder a bit, but I'll be darned if I dont get lost every time I go browse one of those electronics-component mega-sites.

I'm very interested in seeing the pictures, and will be more than happy to host them if you would like to email them to me at [email protected]

Look foreward to hearing/seeing more...

--BillyM

Panelhead · « **Reply #5 on:** 20 Feb 2005, 06:11 pm »

Quote from: BillyM

What version of the 5066 do you have? Also, would you be so kind as to toss up a part-number and source for your upgrade components? I can enjoy music, I can solder a bit, but I'll be darned if I dont get lost every time I go browse one of those electronics-component mega-sites.

I'm very interested in seeing the pictures, and will be more than happy to host them if you would like to email them to me at [email protected]

Look foreward to hearing/seeing more...

--BillyM

Billy,
I think all mine are later version. But the two worked on have different factory chokes in the output. Maybe be one is later version than the other. Have a couple of stock ones, to go with the factory triangular speakers.
It would be easier to locate the parts referenced if I had listed the correct supplier. Mouser is who stocks the 10 uH, 1500 ma chokes used. The part number is 542-4622. They are 1.32 each. The input caps I used are not optimal, going to try 2.2 ufd in the smaller pachage. I am getting a pop on power up, hope the smaller value will tame this. The smaller size will also fit on the pads better.
Murata has the GRM21BR71A225KA01 and TDK has a C20112X7R1C225K that look to be good input coupling cap options. The AVX currently used seems fine. All are available at www.mouser.com
There are some nicer looking SMD out there, C0G/NP0 and the fancy low ESR Murata GQM series. But they are not available in the large values needed for coupling applications.
I also replaced the factory 0.15 ufd caps across the speaker terminals with some molded multlayer ceramics of the same value. These are better for RFI decoupling than poly caps. This value is not the value listed in the Trpiath appl notes. They show 0.47 ufd to ground on each leg for 4 ohm and 0.22 to ground for 8 ohm.
Tried 0.1 ufd poly here for my 15 ohm speakers, but could not tell much so went back to 0.15 ceramics for now.

George

dave_c · « **Reply #6 on:** 20 Feb 2005, 06:30 pm »

Quote from: dwk

I assume these are air-core chokes?

I got into digital amps with the Panny units, and am still baffled by the fact that folks spent time on wire and speaker jacks while the output inductors were ignored - good air-core inductors (perhaps with some shielding improvements to mitigate any extra rfi) would seem to be a no-brainer, aside from the mechanical/space problems.

Steve Nugent replaced the inductors when he modded my Panny. It was a huge difference from stock in all aspects of sound.

BillyM · « **Reply #7 on:** 20 Feb 2005, 11:19 pm »

George, thanks for the photos and the more detailed explaination. I went ahead and resized your images and hosted them. I dont believe you will mind, so I am going to post their links here. If I am over-stepping my bounds, just email me and I'll take them down.

--BillyM

http://webpages.charter.net/wackskier/5066/george1.jpg
http://webpages.charter.net/wackskier/5066/george2.jpg
http://webpages.charter.net/wackskier/5066/george3.jpg

Panelhead · « **Reply #8 on:** 21 Feb 2005, 12:34 am »

Thanks Billy,
Some have had questions about the chokes used. A picture is better than my description. Hope this helps if anyone is looking to try the J.W.Miller chokes.

George

Gary P · « **Reply #9 on:** 21 Feb 2005, 02:06 am »

Thanks Panelhead,

I'll get some of the J.W.Miller chokes to try. Looks like they will fit into the stock case, barely.

Gary

doggie · « **Reply #10 on:** 25 Feb 2005, 01:41 pm »

Quote from: Panelhead

Billy,
I think all mine are later version. But the two worked on have different factory chokes in the output. Maybe be one is later version than the other. Have a couple of stock ones, to go with the factory triangular speakers.
It would be easier to locate the parts referenced if I had listed the correct supplier. Mouser is who stocks the 10 uH, 1500 ma chokes used. The part number is 542-4622. They are 1.32 each. The input caps I used are not optimal, going to try 2.2 ufd in the smaller pachage. I a ...

Thanks for the ecellent tip. I went to Mouser and looked up the part number. They show an 11 week backorder time

Any ideas for a substitute or an alternative source?

Thanks,

Paul

Panelhead · « **Reply #11 on:** 25 Feb 2005, 01:57 pm »

Quote from: doggie

Thanks for the ecellent tip. I went to Mouser and looked up the part number. They show an 11 week backorder time

Any ideas for a substitute or an alternative source?

Thanks,

Paul

Paul,
Email your address. A local store has them in stock for 1.00 each. I can pick some up and mail them to you. If you do not mind going off the board with the chokes, parts express has some 10 uH aircore inductors for 2.19 each. They are rated 300 watts, may be pretty large. My next 5066 project is going to use some 15 uH Solo aircore I picked up. They are large also, but are manageable.
I did not check the availability at Mouser. The reason I looked them up is to double check the ratings. My meter is not repeating measurements very well at this low value of inductance. I can measure one three times and get three different readings. The scale is 2 mH, guess this 0.01 mH is just below the level of resolution.
Good Luck!
George

Panelhead · « **Reply #12 on:** 25 Feb 2005, 03:18 pm »

Quote from: Panelhead

Paul,
Email your address. A local store has them in stock for 1.00 each. I can pick some up and mail them to you. If you do not mind going off the board with the chokes, parts express has some 10 uH aircore inductors for 2.19 each. They are rated 300 watts, may be pretty large. My next 5066 project is going to use some 15 uH Solo aircore I picked up. They are large also, but are manageable.
I did not check the availability at Mouser. The reason I looked them up is to double check the ratings. My meter ...

Thinking about this, it may have been MCM that listed the 10 uH inductors. A catalog came to the house with a chart of inductors. Cannot remember which. Seems like they were 1.99 each if 10 or more are ordered.

doggie · « **Reply #13 on:** 25 Feb 2005, 03:22 pm »

Quote from: Panelhead

Thinking about this, it may have been MCM that listed the 10 uH inductors. A catalog came to the house with a chart of inductors. Cannot remember which. Seems like they were 1.99 each if 10 or more are ordered.

Thanks a lot. I will check MCM and then may take you up on your kind offer.

doggie · « **Reply #14 on:** 25 Feb 2005, 04:11 pm »

I checked back at Mouser and the chokes now show in-stck. Go figure...

eweitzman · « **Reply #15 on:** 1 Mar 2005, 08:56 am »

Moving the chokes off the board will increase the RF emissions from the amp. I hope this doesn't interfere with your other gear or your neighbors'.

- Eric

Panelhead · « **Reply #16 on:** 1 Mar 2005, 11:23 am »

Quote from: eweitzman

Moving the chokes off the board will increase the RF emissions from the amp. I hope this doesn't interfere with your other gear or your neighbors'.

- Eric

Eric,
I have a nearby tuner and phono stage and so far have not noticed any issues. The reworked chip is in a heavy cast aluminum box. Hope it is providing the needed shielding.
I was wondering about the use of larger chokes. But I was more concerned with the choke picking up the noise coming off the chip. I do not have any meters to measure this high frequency trash.
The larger chokes can be thought of as antennas, do you think that they will radiate a lot more RF1 than the stock small ones? They are mounted almost perpendicular to each other in the current board. But my next project is has completely off board air cored inductors for the channel used. The unused channels have stock stocks and terminating resistors.
Had planned on mounting the chokes close together and shielding the chips to reduce the radiation pickup in the chokes. This looks to be easier than shielding the chokes.

George

eweitzman · « **Reply #17 on:** 1 Mar 2005, 06:16 pm »

The RF signal (along with the audio signal) is conducted from the chip to the choke along the wiring between them. In the normal situation, this signal travels on the PCB trace between chip and choke. Tripath's design notes say that this path must be as short as possible, since it's this path that acts as an antenna and transmits the RF to the surrounding space. If you use a large choke that has to be placed off the PCB and farther away from the chip, you've made the path longer. This effectively makes the RF transmitter's antenna longer, resulting in more radiated RF noise.

Even if your box is an effective shield preventing the RF from radiating outwards, the increased RF noise will be picked up by the internal components and outgoing wires. It will go back to the AC and other components along the power cord, back up the interconnects to your source, and out to the speakers. These will also radiate the RF in addition to whatever problems they cause for the components on the other ends. One modifier recommends placing ferrite chokes on all these wires inside the amp.

- Eric