MyRefC build guide - Page 4 - diyAudio
Go Back   Home > Forums > Amplifiers > Chip Amps

Chip Amps Amplifiers based on integrated circuits

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
Reply
 
Thread Tools Search this Thread
Old 28th May 2010, 02:10 AM   #31
Bill_P is offline Bill_P  United States
diyAudio Member
 
Join Date: Sep 2009
Location: New York, the vampire state
Tom, you raised several issues and there is no short answer to any of it. I work for a major corporation designing electronic assemblies (not audio). A given design might be sold a a rate of 50000 a year and a field failure rate of 5% is more than enough to call a halt to production. So if 95% of the units work, I have failed and will be called before an executive review board to explain myself. Obviously this is highly unpleasant and it is far better to design and test thoroughly beforehand. Is it reasonable to apply these standards to the audio hobby? Maybe not, but for me it's ingrained I guess.

The relay circuit uses several components outside the manufacturer's specifications. Even so, given the use of quality components with built in design margins, some or most of the units may work. But it is poor practice and not acceptable in my world. Regarding the diodes, it is not a case of purposefully selecting a weak component. There are fuses and other circuit protectors for that. I have measured samples of 1N4001's on a curve tracer and the breakdown varies all over the map. Some manufacturers seem to label some lots 1N4001 when they actually measure as if they are rebranded 1N4004 types with a 400V breakdown. The economics of producing one part that can be branded with several part numbers may dictate that decision. Regardless, the user can't know how much above the rated voltage a given 1N4001 will tolerate.

The LM3886 mounting issue is a question of mechanical force transmission once the screw is tightened to hold the chip against the heatsink. The screw wants to force the chip flat against the heatsink. The board sticks out further and hits the heatsink and tries to prevent the chip from flattening against the heatsink. These opposing forces put strain on the solder joints at the PC board and on the chip package. If the metal tab on the chip bends even microscopically, it can crack the chip die. Or the strain can cause the solder joints to fail over time. Or the chip may not achieve good contact with the heatsink since the lower part of the chip package will tend to pull away from the heatsink. In that case the thermal path from chip to heatsink will be poor and the die will run hot. In no instance is it an advantage to have the board touch the heatsink.

I removed more material from the Caddock heatsink on the LM318 side because I used an IC socket and do plan to try several different LM318's when the amp is up and running. Getting access to the socket to change op amps required a bit more space than if I had just soldered the LM318 to the board directly.

I will report my impressions about how it sounds but it's going to take a while to finish building it. Before it is installed in the system there are lots of bench tests I want to conduct. The amplifier circuit is really very complex with several nested feedback loops. No decent SPICE model exists for the LM3886 so the simulation part will have to be skipped. The wide range of values for R42 (see the PC board silk screen) is of some concern and I want to get a better understanding of how R42 affects the amplifier stability.
Attached Images
File Type: jpg myref1.JPG (138.9 KB, 1033 views)
File Type: jpg myref2.JPG (116.6 KB, 996 views)
File Type: jpg myref3.JPG (94.6 KB, 963 views)
File Type: jpg myref4.JPG (107.6 KB, 949 views)
  Reply With Quote
Old 28th May 2010, 04:02 AM   #32
diyAudio Member
 
Join Date: Sep 2007
Thanks for the thorough explanation. Your professionalism is readily apparent, and all your comments make sense from that perspective. I'd like to see the quality of your solder joints! The completed amp certainly looks tidy on top.

In the future, please keep in mind that most of us, if not all, are rank amateurs, and we can hardly understand the workings of this thing, much less complicated circuit or component analysis. I look forward to your examination of the completed amp, but bring it down a notch, if you could, so we can all appreciate your hard work. I am eager to learn, but so far I'm just baffled by some of your contribution. It is pretty cool, though.

I'm sure your willingness to share your vast knowledge and insight will be an asset to the rest of the group. Me, I'm happy when sweet music and no magic smoke comes out!

Peace,
Tom E
  Reply With Quote
Old 28th May 2010, 10:44 AM   #33
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Quote:
Originally Posted by madisonears View Post
One thing really bothers me: if so much is wrong with the protection circuit, how come so many amps work just fine?
it is down a different philosophy.
I regularly run relays at 50% to 70% of rated voltage. But that needs the cap to give the relay a kick and ensure it pulls in hard and then relaxes,
That very well researched paper is based solely on the statement that the relay may have poorer conductance across the contacts if not run at rated voltage.

One has to decide whether to run the relay at rated voltage or at some reduced voltage.
I will be running it at reduced voltage. Possibly as low as 50% (12V) but I will also look at drop out time. It seems to me many of the RC timers are not well chosen.

I did appreciate all that time sequence research. It confirmed many ideas I have on the way timed relays operate, but I do it all (very slowly) by hand calculation.
__________________
regards Andrew T.

Last edited by AndrewT; 28th May 2010 at 10:49 AM.
  Reply With Quote
Old 28th May 2010, 11:54 AM   #34
Bill_P is offline Bill_P  United States
diyAudio Member
 
Join Date: Sep 2009
Location: New York, the vampire state
Quote:
Originally Posted by AndrewT View Post
it is down a different philosophy.
I regularly run relays at 50% to 70% of rated voltage. But that needs the cap to give the relay a kick and ensure it pulls in hard and then relaxes,
That very well researched paper is based solely on the statement that the relay may have poorer conductance across the contacts if not run at rated voltage.

One has to decide whether to run the relay at rated voltage or at some reduced voltage.
I will be running it at reduced voltage. Possibly as low as 50% (12V) but I will also look at drop out time. It seems to me many of the RC timers are not well chosen.

I did appreciate all that time sequence research. It confirmed many ideas I have on the way timed relays operate, but I do it all (very slowly) by hand calculation.
Just curious - why will you be running reduced relay coil voltage?

The timers are not precision circuits since they depend on transistor parameters like Vbe, Ib, and Hfe for thresholding. All the parameters change significantly with temperature and that affects timing too. The timer must not respond too quickly or it may cause relay drop-out when playing music at full power. What problems do you see with the time response of the circuit as originally designed?
  Reply With Quote
Old 28th May 2010, 12:11 PM   #35
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
running reduced voltage (long term) has the advantages of :
1.) drawing less current from the supply.
2.) relay running at lower temperature.
3.) relay drops out more quickly after loss of supply.

The 40ms delay before the relay loses it's supply is absurdly long.
<10ms would be much better to prevent bangs, cracks, pops and clicks getting to the speaker in event of mains failure.
The 220k and 100uF giving RC = 22seconds can be improved. It does not need to be that long under any circumstances.
__________________
regards Andrew T.
  Reply With Quote
Old 28th May 2010, 02:39 PM   #36
Bill_P is offline Bill_P  United States
diyAudio Member
 
Join Date: Sep 2009
Location: New York, the vampire state
Andrew,

Just a few comments and results of measurements here.

The relay tested with one amp DC through the contact measures 4.4 milliohms resistance (Kelvin 4 wire connection). Initially there was a transient change in the resistance when I lowered the coil voltage. After a minute or so it settled out and the resistance returned to its original value as measured at 24VDC across the coil. Letting the relay run for a half hour, the transient no longer occurred and the contact resistance did not change with coil voltage. 14.7 volts was needed to pick the relay and it dropped out at 4.1 volts. This is a sample of one at room temperature.

In my experience a pulse as short as 1 millisecond can be heard as a click through a speaker. The relay mechanical response time is specified as 5 milliseconds max so I doubt it will be possible to eliminate all speaker noise in the event of a transient.

Mains at 50 Hz will have a period of 20 milliseconds. If your objective is to drop the relay within 1/2 line cycle, that will be quite difficult and the circuit might misfire under small line disturbances that would otherwise not disrupt the amplifier.

The discussion about the microscopic detail of the relay and its driver is probably off topic for a general build thread and may not interest many members who bought the kit. If you have ideas for improving the protection circuit, I am interested in your ideas and will help with the investigation if I can. Maybe a new thread or direct PM would be a better venue for further discussion.
  Reply With Quote
Old 28th May 2010, 03:24 PM   #37
diyAudio Member
 
regiregi22's Avatar
 
Join Date: May 2009
Location: Madrid (Spain)
Bill,
-I didn't knew that, why C17, C18, C19 and C20 are not being used and supplied with the kit?

-Why are for the clamping diodes you placed under the board? What are they?

-I didn't knew about the problem of lack of supply for the relay. Is that a design issue from the beginning? Never heard of it...

-Neither knew that an 1N4001 is too low rated for this position.


Uriah, did you have the same problem as Bill when placing the LM3886 on the heatsink?


Andrew, is there any minimal disadvantage in running the relay at lower voltage than recommended?
I am waiting too for your thoughts of the RC timers, why they don't seem addecuate.
__________________
diyAudio, doing it as big as you can, JUST BECAUSE WE CAN!
  Reply With Quote
Old 28th May 2010, 03:57 PM   #38
Bill_P is offline Bill_P  United States
diyAudio Member
 
Join Date: Sep 2009
Location: New York, the vampire state
Quote:
Originally Posted by regiregi22 View Post
Bill,
-I didn't knew that, why C17, C18, C19 and C20 are not being used and supplied with the kit?

-Why are for the clamping diodes you placed under the board? What are they?

-I didn't knew about the problem of lack of supply for the relay. Is that a design issue from the beginning? Never heard of it...

-Neither knew that an 1N4001 is too low rated for this position.


Uriah, did you have the same problem as Bill when placing the LM3886 on the heatsink?


Andrew, is there any minimal disadvantage in running the relay at lower voltage than recommended?
I am waiting too for your thoughts of the RC timers, why they don't seem addecuate.
Some people think C17-C20 hurt performance rather than help. There have been discussions about this on the forum and maybe in the original group buy thread.

The clamping diodes are across the capacitors in the relay circuit. The explanation is in the PDF I posted.

The low supply voltage for the relay has evidently been part of the design from the beginning judging from the silk screen on the PC board showing component values.

The 1N4001 is rated too low and must have been a design oversight. Easy enough to change for those who are concerned about it.

My measurements indicate there may not be much disadvantage to undervolting the relay in the typical case. My own preference is to run near rated voltage and I have changed R14 to do that.
  Reply With Quote
Old 28th May 2010, 04:19 PM   #39
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
I agree with BillP.
The circuit as designed and in the PCB works.
The improvement offered by BillP is optional.
Whether one implements the improvement is down to whether or not one requires the speaker isolating relay to operate with a long term voltage of 100% of rated voltage or 62% of rated voltage.

The builder has to weigh up the pros and cons and decide.

AS IS, WILL WORK.
__________________
regards Andrew T.
  Reply With Quote
Old 28th May 2010, 04:34 PM   #40
diyAudio Member
 
regiregi22's Avatar
 
Join Date: May 2009
Location: Madrid (Spain)
What are the cons of supplying it with a lower voltage?
__________________
diyAudio, doing it as big as you can, JUST BECAUSE WE CAN!
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
New guide from TI FastEddy Parts 0 19th August 2007 08:36 PM
Various questions from a diy noob regarding the myrefC. thegeek Chip Amps 8 7th June 2007 04:49 PM
Everything required to assemble MyRefC and Kookabura. Tony D. Swap Meet 3 18th January 2006 02:54 AM
what is a wave guide? Joules Multi-Way 7 12th May 2005 02:25 PM


New To Site? Need Help?

All times are GMT. The time now is 10:14 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2