National Semiconductor LM series & Ti LM series differences that affect circuit funct
Hi i"m new to this but want to start by hard wiring an LM1875. Super simple and as I have had jewellery as a hobby for 35 years, fiddly & small is the way I roll. One issue that nags me is this. Most of the circuit designns that interest me (Chip Amp. Audio Sector and 47labs)
were all designed with the Natioall chip. The Texas Instruments specs seem very close but not exactly the same. Are there any differences that affect the above mentioned circuits and so how? The only mention of this I have found is how the change affected design software
like spice. Way over my head .Did TI change anything or "improve " anything?
Can someone please clear this question up for me? Am I creating much ado about nothing and if so prove it?
Cheers theaudiopath.
.
Hi i"m new to this but want to start by hard wiring an LM1875. Super simple and as I have had jewellery as a hobby for 35 years, fiddly & small is the way I roll. One issue that nags me is this. Most of the circuit designns that interest me (Chip Amp. Audio Sector and 47labs)
were all designed with the Natioall chip. The Texas Instruments specs seem very close but not exactly the same. Are there any differences that affect the above mentioned circuits and so how? The only mention of this I have found is how the change affected design software
like spice. Way over my head .Did TI change anything or "improve " anything?
Can someone please clear this question up for me? Am I creating much ado about nothing and if so prove it?
Cheers theaudiopath.
.
Thanks jackiinnj. That is standard corporate OP in anticipation of takeover. No surprises there.
What I am actually seeking is those who have tried both chips in the same circuit and found an audible difference + or -. I Realise it is AM in your neck of the woods but 3pm here. Also where can I find matched and burnt in chips so I have even a remote chance of getting 2 stereo channels of the amp to match.
Cheers Mark.
What I am actually seeking is those who have tried both chips in the same circuit and found an audible difference + or -. I Realise it is AM in your neck of the woods but 3pm here. Also where can I find matched and burnt in chips so I have even a remote chance of getting 2 stereo channels of the amp to match.
Cheers Mark.
Mr. Path,
Why do you assume that there is any difference in the National power amp chips before and after the TI acquisition? I'm sure that if you ask them, you'll find that they are made in the same wafer fabs, assembled and tested at the same factories, and all this done with the same equipment and (likely) many of the same people as pre-acquistion. If there is any substantive (i.e., non-cosmetic) change in the datasheet, it's likely a result of normal revision process (find an error and correct it). Semiconductor manufacturing is a highly complex process. You don't just pick it up and move to another 'interchangeable' factory like a clothing manufacturer would do with a pair of blue jeans.
I would expect any two chips from the same wafer lot to perform substantially identically to each other. Even with two random units of the same product, the marvel of negative feedback is that the two will likely perform 'identically' in the final application circuit.
Why do you assume that there is any difference in the National power amp chips before and after the TI acquisition? I'm sure that if you ask them, you'll find that they are made in the same wafer fabs, assembled and tested at the same factories, and all this done with the same equipment and (likely) many of the same people as pre-acquistion. If there is any substantive (i.e., non-cosmetic) change in the datasheet, it's likely a result of normal revision process (find an error and correct it). Semiconductor manufacturing is a highly complex process. You don't just pick it up and move to another 'interchangeable' factory like a clothing manufacturer would do with a pair of blue jeans.
I would expect any two chips from the same wafer lot to perform substantially identically to each other. Even with two random units of the same product, the marvel of negative feedback is that the two will likely perform 'identically' in the final application circuit.
Dear Mr L
I can understand that you would assume that I am making that assumption. But you are replying to one assumption with another. YES this is the most likely situation but are you guessing or do actually
KNOW? Assumptons can be dangerous. According to the meticulous measurement of the LM chips by Peter Daniel they of the same batch can vary significantly more than the 10-30 mV DC offset recommended by Chip Amp as optimum in the instruction PDF for the Lm1875 kit.
Cheers Theaudiopath
I can understand that you would assume that I am making that assumption. But you are replying to one assumption with another. YES this is the most likely situation but are you guessing or do actually
KNOW? Assumptons can be dangerous. According to the meticulous measurement of the LM chips by Peter Daniel they of the same batch can vary significantly more than the 10-30 mV DC offset recommended by Chip Amp as optimum in the instruction PDF for the Lm1875 kit.
Cheers Theaudiopath
Peter Daniel's implementations are usually DC coupled.
These implementations are very sensitive to small tolerances resulting in big variations in output offset.
You have choices
a.)
AC couple your Power amplifiers
b.)
DC couple your Power amplifiers and include set up adjustments to minimise output offset
c.)
DC couple and fit all the correction, detection and isolation gear to protect your speakers.
d.)
DC couple and take no precautions to protect your speakers.
Note that none of the 4 options blame the manufacturer for the varying offset problem.
These implementations are very sensitive to small tolerances resulting in big variations in output offset.
You have choices
a.)
AC couple your Power amplifiers
b.)
DC couple your Power amplifiers and include set up adjustments to minimise output offset
c.)
DC couple and fit all the correction, detection and isolation gear to protect your speakers.
d.)
DC couple and take no precautions to protect your speakers.
Note that none of the 4 options blame the manufacturer for the varying offset problem.
Last edited:
Cheers AndrewT & Wintermute for your replies. I Think purchasing a few Chips from Element 14 is a good starting point. @ $3 & a bit each makes for cheap as chips (sad pun intended) experimentation.
Have most of the hardware including cases aside from the resistors & some of the caps. A Pair of paralleled 12v mobile phone relay station back up batteries per channel should supply ample clean power. I Think following the popularity of all Mr Daniel's amp incarnations I'll start there. Why mess with success.
What sort of trim pot would you suggest for this project?
Cheers Mark.
Have most of the hardware including cases aside from the resistors & some of the caps. A Pair of paralleled 12v mobile phone relay station back up batteries per channel should supply ample clean power. I Think following the popularity of all Mr Daniel's amp incarnations I'll start there. Why mess with success.
What sort of trim pot would you suggest for this project?
Cheers Mark.
You did say hardwiring, right?
I would suggest this type of trimpot (3006P) rather than the usual miniature square format types which have wobbly leads that fracture easily (or the solder does) if not pulled down flat on a PCB. If you need to screw them about more than a few times, the benefit of a PCB with 0.5 mm mounting wires becomes more obvious. Both types are stocked by au.E14.
The difference with this type, IIRC, is that leads emerge directly from epoxy sealant which makes for a much more rigid and easier to make connection, as I'm sure you're able to do to perfection. This should wind up as good as a board mounted type and of course, you can orient it in the most accessible direction.
Last edited:
Thanks Ian Fiinch
I actually intend to use no bread board at all. Just a block of wood with chiseled & pre-drilled holes.
Mount the components upside down & secured with glue & hard wired with spare silver wire from the
top. Your reply for some reason made me think about another issue. I noticed there seemed to be a disprorational no of issues with the LM1875 component. I Find
this quite disturbing.
As I have no idea if it is due to quality control issues, heat stress while soldering 5 points dissipating through such a tiny chip or bad technique by the DIYers fake chipsets or all of the above
I actually intend to use no bread board at all. Just a block of wood with chiseled & pre-drilled holes.
Mount the components upside down & secured with glue & hard wired with spare silver wire from the
top. Your reply for some reason made me think about another issue. I noticed there seemed to be a disprorational no of issues with the LM1875 component. I Find
this quite disturbing.
As I have no idea if it is due to quality control issues, heat stress while soldering 5 points dissipating through such a tiny chip or bad technique by the DIYers fake chipsets or all of the above
Thankyou infinia
Your reply is most "helpful" I am not buying a kit but hardwiring from scratch. But you are correct that some of the best chips never make it into our hands. Had one given to me for a class A output stage on a HK 970 CD player twice the output clean of the advertised spec. Have forgotten the model but not the sound. Quality Control nab the very best, wholesalers and kit suppliers get second best and we who are left have to suck the excrement from the corporate sphincter. I am being overly controversial & cynical in the vain remote hope of being proven wrong.
Your reply is most "helpful" I am not buying a kit but hardwiring from scratch. But you are correct that some of the best chips never make it into our hands. Had one given to me for a class A output stage on a HK 970 CD player twice the output clean of the advertised spec. Have forgotten the model but not the sound. Quality Control nab the very best, wholesalers and kit suppliers get second best and we who are left have to suck the excrement from the corporate sphincter. I am being overly controversial & cynical in the vain remote hope of being proven wrong.
I can't prove you wrong, but unless the chip manufacturers bin their ICs such as is done with LED chips, I'm not buying in to this.
One of my favorite stories about how things worked went something like this. It mainly applied to things like memory chips back in the day with different speed grades but was broadly applicable to other aspects of specifications.
Nobody deliberately tried to make slower chips or continued to do so as manufacturing techniques improved. What would happen is: The chips were tested as they came off the lines and manufacturing was continued until enough of the high speed grade to satisfy orders were made. If some chips exceeded specs, the manufacturer didn't care as they were only worried about the minimal testing to fill orders. If there were orders for slower chips, they would just keep the line running, testing at this lower specification and/or retesting chips that failed the higher speed specification. Once again, even if the chips were wildly better then the speed grade they were being sold at, they didn't care.
So during the RAM shortage of the 80's there were a lot of shops that would retest chips. Manufacturer rejects, dumpster diving, chips sold at lower speed ratings. Just a person sitting at a work bench loading RAM chips into a test fixture to see what they actually ran at. These would be 'cherry picked' to find the best ones for use in desktops. Indeed it kept a lot of small shops going during that time.
I recall telling one of my managers/supervisors at the time, "The batch came out bad." He corrected me and said "Some of the products we make are better then some of the other products we make, but we don't make bad products."
It still makes sense to me. I mean there are the TDA2030, TDA2040, TDA2050. Why bother making the TDA2030? If you have a manufacturing process to make the better grade pin compatible chips, why bother making the lousy ones? Line change overs are tedious and costly.
I did a google search to see if anyone had pictures of TDA2030-50 dies and came up empty. I may have a couple floating around. It may have some value to take a couple apart for a look see.
You're creating much ado about nothing...
I guarantee you that the designs did not change. The masks used for producing the chips did not change. It is actually very expensive to make a datasheet change on a product in production - key customers need to be notified (big deal!), distributor stock may have to be called back and scrapped, etc. - so I doubt the datasheets changed aside from changing National Semiconductor to Texas Instruments. If you have specific examples and can show both datasheets, I'm all ears. Changing the design (even a small tweak) is even more expensive. The part has to be re-characterized and the datasheet updated (with all those consequences). The production line hasn't changed. The parts are the same. They just have a TI logo rather than the National logo.
The biggest change was to the application notes where the authors' names were removed.
~Tom
That's called binning. It's hardly used anymore as it is rather expensive to do. That's not to say it is never done. It's just rarely done. Only if the business case justifies binning is binning done. The rest of the time we just rely on building kick-a** circuits that make the customers happy...
~Tom
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.