The OFFICIAL MSM5540RS/M5540 Sansui Gx700 40pin IC Thread! NUDIES

[kevzep]

The typical failure mode of the MSM5540 is to generate "garbage" on the display.

Not necessarily, it is a common fault with the 5540 to not get Quartz Lock as well. I have had this more than three times in the last year, and when people are telling me its the Crystal.
I work on a lot of these Receivers and I have never had a faulty crystal. I am not saying they don't fault, but its not as common as what you would think reading around this forum.
The 5540 doesn't generate "garbage", they either have segments which are ON all the time, or they have segments that are off all the time (missing segments). At certain positions it looks a little weird but the chip does not "generate" these characters, the chip is the driver that turns segments on and off depending on the input information, the chip is not a micro-processor.
The other thing it does, which is not exactly clear how at this point, is it has something to do with the quartz lock....Further investigation is required.

 

[Sansuiman]

Understood, but the 5540 can be said to be generating garbage to the display by virtue of the fact that it isn't correctly controlling segments at the VFD in response to data at the input side. I've been lucky - apparently - not to have any of my gear that contains an MSM5540 IC fail in the manner frequently posted about here. That being the case, I've not engaged in troubleshooting this setup myself.

I figure the display data input is probably BCD or some such scheme, and there is a shift register within the MSM5540 to convert that into multiplexed segment/digit control of the display on the output side of the chip.

It is only a 40 pin device, so not enough discrete I/O lines for direct control of every segment and digit. VFD's (and most digital displays) of any complexity are almost always driven by a multiplexed display driver anyway. If there are no missing or locked on segments presents and the display is reading out just at an incorrect range, then I consider it unlikely the MSM5540 is the root issue.

Upon further testing, my theory in that regard may prove incorrect, but is still within the realm of possibility as it stands presently.

The quartz lock functionality is a secondary issue, and if not working can certainly be an internal fault of this chip, and possibly extant indepent of the normally seen display driver problems. I'll be following this thread to see the outcome.

Has anyone pulled the dataheet for the MSM5540? Guess I'll go digging for it. Eventually, a MCU of some sort will need to be employed to replicate it's functionality, or any component using it will be a boat anchor upon the failure of this chip, as parts units to cannibalize will eventually be depleted. The stock of T-80's is already being hit now to save better tuners and receivers.

 

[kevzep]

Has anyone pulled the dataheet for the MSM5540?

If you find one, let us know, I have been periodically looking for years.....as have others with no success.

Whatever it is that fails in the Quartz lock circuit, most definitely has something to do with this chip.

I have two faulty 5540's here, but I have little to no time to look at this until concert season is over, so it wont be for a month or two.

 

[Hyperion]

If you find one, let us know, I have been periodically looking for years.....as have others with no success.

Darn right, I haven't been able to find one either.
We've had a couple of people now who have posted up that they may be able to put together an equivalent using more modern IC's - BUT understandably they need to know more about what the 5540 is doing, getting a data sheet would solve this part of the puzzle...

 

[Sansuiman]

A quick note, some research has revealed that Oki sold their semiconductor division to Rohm in 2008, and the unit now operates under the name Lapis Semiconductor. A link on their site to the obsolete/ EOL product catalog states that an inquiry needs to be made directly to their service department as this data is no longer online.

I will attempt to work my way through the system and see if a datasheet can be procured. If I can get that far i will also inquire if there are separate data sheets for MSM5540 and MSM5540RS, which apparently is/was a pin compatible variant.

 

[kevzep]

A quick note, some research has revealed that Oki sold their semiconductor division to Rohm in 2008, and the unit now operates under the name Lapis Semiconductor. A link on their site to the obsolete/ EOL product catalog states that an inquiry needs to be made directly to their service department as this data is no longer online.

I will attempt to work my way through the system and see if a datasheet can be procured. If I can get that far i will also inquire if there are separate data sheets for MSM5540 and MSM5540RS, which apparently is/was a pin compatible variant.

Excellent, I am glad someone has the time to follow it up, that would be great if we could obtain the datasheet.

Nice detective work...

 

[johnhoef]

I received the new crystal, and installed it this morning. Sadly, no change. Not sure what to do next. There is a trim cap in the crystal circuit. Do I dare adjust that? I would think not. If the crystal really was the problem, I would see SOME change.....none at all. I'm beginning to think replacing the MSM5540 is the answer? The prescaler output feeds directly to pin 39 of the MSM5540, so the rest of the logic takes place inside that chip?

 

[Sansuiman]

That's certainly a different failure mode than what we are accustomed to seeing with a bad MSM5540 (missing/locked on segments and loss of quartz lock), and would be the first documented instance of this particular symptom that I'm aware of.

I'm not sure what else to suggest, short of sourcing a replacement chip. Still seems awfully strange to me that it could fail internally in such a fashion, as it is still functional, just shifting the displayed frequency range by what appears to be a particular amount of offset.

 

[johnhoef]

Sure would appreciate any further advice as to the next step here.....replace the chip?

 

[Hyperion]

Perhaps something up with the prescaler? - but that is rather less likely to be faulty than the main IC.

 

[johnhoef]

As mentioned above, I have verified the prescaler does indeed divide it's input by 100, converting the mhz from the local oscillator to a khz frequency, and supplying that signal to pin 39 of the big chip. I would think there is nothing more to do with the prescaler.

 

[Sansuiman]

Just to be sure I'm clear on the present situation, the quartz lock is working and the display reads out legible (no segments locked on or off) just the displayed range of frequency is shifted incorrectly, right? Makes me wonder if the prescaler is not outputting in the correct range.

Obviously, the MSM5540 will decode and drive the display based upon the frequency input at pin 39 (output from prescaler) and will not have anyway of "knowing" that the input data is not in the correct range - if it is in fact wrong - unless the prescaler output is beyond the acceptable input limits of the 5540. Without a datasheet, we don't know what that range is as I don't recall the service manual for the TU-719 and 919 (which I have) detailing that info.

Do you have a function generator that can be setup to spoof the input of the 5540 and see if supplying a different range of frequency to pin 39 will render output at the display in a correct range of 88.1-107.9Mhz for FM? Just a thought.

 

[kevzep]

the MSM5540 will decode and drive the display based upon the frequency input at pin 39

Yeah I think that is where the issue is in the chip.

Pretty much all the faults I have had with segments, quartz lock, have been the chip.

 

[johnhoef]

I am still not 100% confident, and would like to know more about how this pre scaler works. Studying the schematic, and probing with my scope, I find the AM input turns on and off as the selector switch is turned. However, the FM input is always on. I expected it to turn off in AM mode. Scoping the output, I do see a change. When switching to AM, the pre scaler output changes to a higher frequency. This makes sense since the numerical display shows a higher number. Think I'll try disconnecting pin 39, and injecting an unmodulated signal from my RF generator into pin 39. I expect to see the digital display to change as I change the RF frequency. I'll let you all know the result. I'm curious to see how the digital display corresponds. There are MSM5540 chips available at $40 a shot...a bit expensive considering this is somewhat of a shot in the dark.

 

[Hyperion]

and injecting an unmodulated signal from my RF generator into pin 39.

Not a bad idea - be careful with the input level.

 

[kevzep]

There are MSM5540 chips available at $40 a shot...a bit expensive considering this is somewhat of a shot in the dark

Really? Where from? As far as I can tell and others, you cannot find them any more. Last one I paid $65US and that was 10 years ago..

 

[Sansuiman]

Ceitron has them back in stock, through their site directly at 29.95 each, min shipping is $12.95 (domestic US). They also have them listed on ebay at 39.90 apiece free shipping (presumably in US only).

I suspect they are ordering these in small lots (maybe 20-50 at a time) from one of a handful of obsolete semi distributors in China. Yeah, I know the concern with Chinese sourced components and proliferation of counterfeits, but lacking other good options I took a gamble on this and ordered a few.

Doing some digging around online reveals that there are a couple of Chinese outfits that have residual stock of under 200 pieces on hand. No retail, and minimum buys involved, so really it falls to an outfit like Ceitron to front the money and stock these for the benefit of guys like us here trying to maintain old equipment.

I've picked up NOS Sansui QS chips (HA1327/1328) in much the same way to have spares on hand through an ebay seller a while back. Funny that they arrived in worn antistat bags marked "SANSUI ELECTRIC CO. LTD." and were clearly reclaimed from Sansui factory stock at some point. 40 years sitting in bags, and someone kept them warehoused that entire time. Remarkable.

I'll vouch for these being legit by the way, just had three NOS MSM5540 arrive from Ceitron yesterday afternoon in an antistat plastic strip. Didn't want to post until I had them in hand, lest I get anyone excited about them being available again prematurely. Get 'em while they're hot.

Here ya go:

http://www.ceitron.com/cart/advance...words=Msm5540&search_in_description=1&x=0&y=0

 

[johnhoef]

As far as counterfeits go, I would be more concerned with output devices that draw a substantial amount of current. In the case of the MSM5540, or any other low power chips, I am less concerned. Sansuiman, thanks for sharing that source with us. I may do the same and pick up a few to keep on hand, and while I'm at it, install one in this unit I am repairing. I will also install a 40 pin socket, just in case I need to swap the chip out again.

 

[Hyperion]

I will also install a 40 pin socket, just in case I need to swap the chip out again.

Ideally you won't want to be doing this too often.

But it would be helpful if someone in this game had a 'mule' system, then they could check suspect 40 pin IC's with minimum difficulty.

 

[mattsd]

Ceitron has them back in stock, through their site directly at 29.95 each, min shipping is $12.95 (domestic US). They also have them listed on ebay at 39.90 apiece free shipping (presumably in US only).

They're real and work great, I got one about 6 months back for repairing a G-9700 display, worked like a charm and the owner says it's still going strong!