MAX1044 causes of damage & alternatives

[soulsonic]

Over at the other forum, some people were complaining that their MAX1044 charge pump chips had failed. Several of them had come to the conclusion that it must be caused by too much voltage as it's only designed to run on 10v. But I honestly don't see how that can be a serious issue if it runs on a 9v battery. I've never measured a 9v to be anywhere near 10v and most of the time, they are considerably less than 9v, so I think the problem lies elsewhere.

Specifically, I believe the chips are failing due to ESD damage. I hypothesize that a likely unprotected part of the where static could damage the chip would be at the battery clip. According to the functional diagram in the datasheet, the V+ pin connects to a MOSFET switch inside the chip that's part of the charge/discharge circuit. Without protection on the V+ input, I believe it could be very easy for damaging static discharges to enter into the chip and damage the MOSFET switches. And I believe that this is what has been happening to people's chips "in the field".

I encourage others here to help enlighten us on this subject. Does this seem like a plausible explanation for the problem? What would you recommend as a good protection scheme? A zener diode?

I tried to bring this concept up at the other forum, but it seems to have been completely ignored (as usual), so I wish to continue the discussion here.

[lmkv15]

HI,

Simply not taking the MAX1044. Takes the LTC1044 or LTC1054 or TCA962 for this. This chips have not Problems.

regards Uwe

[gaussmarkov]

HI,

Simply not taking the MAX1044. Takes the LTC1044 or LTC1054 or TCA962 for this. This chips have not Problems.

regards Uwe

it looks like the MAX1044 is a newer chip than the LTC1044. the MAX1044 datasheet says "They are pin compatible with the industry-standard ICL7660 and LTC1044."

uwe, it sounds like you think the MAX1044 is inferior in a lot of ways. are there conditions under which you might prefer the MAX1044?

cheers, gm

[soulsonic]

The LTC1044 is still a CMOS Device, which means that it is susceptible to ESD damage in the same way as the MAX1044. They are identical as far as this issue is concerned.

[WormBoy]

At the risk of hitting that nerve again .... I build a FSH for a friend of mine, with a MAX1044 in there (TonePad PCB). It worked fine at my place, fine at his place until he took it to a rehearsal, where it died last week. My autopsy confirmed that MAX has passed away. Rounding up the usual suspects ... the FSH was powered by a 1Spot at the time of death. I know this does not prove anything beyond reasonable doubt (might still have been a flaky MAX, an error in the layout, or my soldering), but it does not make me feel entirely confident about the safety of the 1Spot, at least in combination with the MAX ...

[R.G.]

At the risk of hitting that nerve again .... I build a FSH for a friend of mine, with a MAX1044 in there (TonePad PCB). It worked fine at my place, fine at his place until he took it to a rehearsal, where it died last week. My autopsy confirmed that MAX has passed away. Rounding up the usual suspects ... the FSH was powered by a 1Spot at the time of death. I know this does not prove anything beyond reasonable doubt (might still have been a flaky MAX, an error in the layout, or my soldering), but it does not make me feel entirely confident about the safety of the 1Spot, at least in combination with the MAX ...

It's possible. The MAX 1044 has an absolute maximum input voltage of 10.000000V and it dies above that. If something, anything, put more than 10V on it, it properly should have died. My one sensitive area on 1Spots is that the voltage runs a touch high; they're commonly 9.4-9.5V out of the box, intended to be the voltage of fresh alkaline batteries. So with a MAX1044, there's only half a volt or so of headroom before the MAX turns into a pumpkin.

That being said, there is a MAX1044 in every Visual Volume 10th anniversary edition ever made, and we have not yet seen a rash of deaths on the MAX. That doesn't say some combination of a run of high 1Spots and sensitive MAX chips won't do it, but it hasn't yet.

Confidence is a funny thing. You have only one test, so you have to rely on the 100% of your evidence, which is one unit. We sell literally thousands of 1Spots and get a handful of problems, some of which turn out to be 1Spots which have failed (nothing is perfect!) and some of which turn out to be other problems. What I haven't seen yet is anything that looks like an epidemic of 1Spots killing pedals. We each have to go with the numbers we see.

[soulsonic]

I'm convinced that these MAX1044s used in DIY projects are getting killed by static damage. I think static protection on the power input should be mandatory with a 1044 or any other CMOS-based voltage converter. I see that the Tonepad FSH layout does not include this protection.

[jakehop]

Change your MAX1044's to LT1054's - they can put out 10 times the current, and have a max of 15V supply!

I have recommended a lot of 1Spots, and all the guys and gals I know who use them, are very happy with my recommendation. Good job R.G.!

For the DIY'er, check the net for "Switchmode power supply" information, or build one with a transformer. I either design my effects around a specific PSU which I include or built in, or design it for 9,5-10V operation. Better safe than sorry.

Kind regards, Jake

[R.G.]

... Good job R.G.!

Gosh, thanks! But - er -

I should be clear about it - the 1Spot is not my design. It predated me working for Visual Sound. I just have to examine any failures and figuring out what to do if we do have problems. The 1Spot was designed by a power supply design shop that does nothing but design power supplies and these guys are as good as the group I worked in when I as eating, breathing and living power supplies.

It's probably wrong to talk about "the" 1Spot too. There have been a couple of major designs and a few revisions as legislative requirements changed. The most current 1Spots are, for instance, in compliance with California's energy conservation laws which will go into effect this summer, I think, and which will effectively outlaw non-switching wall warts as not being energy efficient enough. It was RoHS compliant on time, etc.

[WormBoy]

We each have to go with the numbers we see.

Well, the statistics are on your side; my n=1 does not really make a case in court But the human brain (mine included) primarily works by extrapolation from personal experience, so you'll have to excuse my less-than-total-confidence in the 1Spot for now while I'm trying to confince my brain to work differently ... The maximum voltage of the MAX is indeed dangerously close to the normal output of a power supply ...

I'm convinced that these MAX1044s used in DIY projects are getting killed by static damage. I think static protection on the power input should be mandatory with a 1044 or any other CMOS-based voltage converter. I see that the Tonepad FSH layout does not include this protection.

The Tonepad layout does include a 220uF cap on the 9V+ that's also powering the MAX ... or do I need a smaller cap to protect against static?

[WormBoy]

Since this thread was sliced out of another thread I don't feel a hesitation anymore to dive just a little deeper into the MAX issue. I did some searches on this and other forums and the most promising/simple protection scheme I found so far is the use of a 9.1V Zener diode. In another thread Bajaman suggested to wire a Zener "across the dc input socket to ground (positive band on zener to positive supply rail input)" with a "100 ohm resistor between the battery + terminal and the zener". This way the Zener acts as shunt when the voltage exceeds approx. 9V.

I also found this thread http://www.freestompboxes.org/viewtopic.php?f=22&t=341&p=3271&hilit=max1044+static#p3271 on the subject, but that obe apparently bled to death.

On DIYstomps there was reference to another solution (if I read correct, on RG's advice) placing a Zener in series with a 10 Ohm resistor in the 9V+ line (so not going to ground). I'm not sure how that works to reduce peak voltages though ... Can anyone shed some light on which solution to use under which circumstances?

There are alternatives to the MAX which can deal with higher supply voltages. Some versions do not have a boost function and may produce audible whine. At this point I am considering to use an LT1054. No boost function, but since it operates on 25kHz already, it should not need one (also delivers more current). However, pin 1 has a different function, and should probably be bent to avoid connection to 9V+ as in the boost function of the MAX.

Another "interesting" little thing: my dead MAX was in a Maestro FSH clone (Tonepad PCB). I replaced it with a 7660 (I thought it was the one with the boost function, but am not entirely sure) and the sample/hold function sounds different ... it was kind of lifeless no matter how I tweaked the trim pots. I extracted a MAX from another pedal I had lying around, placed it in the FSH and the sample/hold as I knew it was back. So it appears that the type of chip can have an effect on the sound, at least in the FSH (maybe because it feeds the noise transistor, and the frequency affects the noise the transistor puts out?).

[Deric]

I've wondered about static with the Max1044. I use a 1 Spot to power around 10 pedals including a PT-80 with a charge pump. I've measure the output of the 1 Spot with a supply voltage of 115VAC - 120VAC and between zero and 10 pedals and always get 9.4VDC. I have also measured the output while connecting and disconnecting other pedals from the chain to see if the voltage spiked. It never does.

I have somehow fried 3 Max1044 chips in my PT-80. The last one popped when I added another pedal to the chain with everything powered up. I've since changed to a 7660S and so far, so good. I haven't noticed any difference in tone but since I have a few 1044's left I'll swap it out and listen when I get a chance.

RG-
Does the VV 10th Anniversary pedal have any other "protection" for the Max1044?

FWIW, I have sold the 1 Sopt at retail for several years and have never heard of one frying a pedal. We did see a very high failure rate with the 1 Spot itself when we first started selling them but they have been extremely reliable for the last few years.

[eggstein]

TC7662BCPA is a good alternative.

[WormBoy]

I've since changed to a 7660S and so far, so good.

But if its static that's killing the chip, than the 15V or so that the 7660 can stand is still not very much. I would suspect that the 1Spot is sometimes producing a little over 9.5V temporarily, and since that voltage is already close to the max of the MAX ... Has anyone fried a MAX with another power supply than the 1Spot? So far, my Boss PSA has not given me problems, but I have only played my MAX-pedals at home, and not under more "extreme" conditions.

I haven't noticed any difference in tone but since I have a few 1044's left I'll swap it out and listen when I get a chance.

I think that the noise transistor in the FSH is a bit special in that respect, but you never know! I will order a few LT1054's and try them in my PT-80. Because they can deliver more current, it might make a difference.

TC7662BCPA is a good alternative.

I have not seen them yet with my suppliers in Europe ... From the specs, they have greater tolerance for input voltage (up to 18V), more current than the MAX (40mA) but not as much as the LT1054 (100mA).

[R.G.]

I've wondered about static with the Max1044. I use a 1 Spot to power around 10 pedals including a PT-80 with a charge pump. I've measure the output of the 1 Spot with a supply voltage of 115VAC - 120VAC and between zero and 10 pedals and always get 9.4VDC. I have also measured the output while connecting and disconnecting other pedals from the chain to see if the voltage spiked. It never does.

That's been my experience. I've tried to set up circumstances for spikes, with sudden on/off cycles, under load, under no load, light load, full load, etc. I've cycled loading on and off instantly with a pulse generator driving a power MOSFET to switch the load on and off; got that to switch in under 100nS which should have driven any inductive or pseudo-inductive dumps crazy, and I've frequency-swept pulse loads on it from DC to well over 250kHz, and nothing shows up acting odd. There's no oveshoot, undershoot, oscillatory behaviour, or - worst! - "turn and run" inductive behaviour that I can find.

This was an outgrowth of the few boutiquers who thought the 1Spot was their frying problem; I characterized as well as I could. There are many high-dollar audio amplifiers that will go south on you with the frequency-swept switched loading. I got that from some stuff I read by Martin Collums; it seems to pretty ruthlessly expose oddities in amplifiers, which is what all power supplies are - just very dull ones.

I have somehow fried 3 Max1044 chips in my PT-80. The last one popped when I added another pedal to the chain with everything powered up. I've since changed to a 7660S and so far, so good. I haven't noticed any difference in tone but since I have a few 1044's left I'll swap it out and listen when I get a chance.

I'd appreciate hearing about anything you find. I do know that the normal inverting configuration for the 1044 leaves pin 6 and 7 open, and possibly prone to static problems.

Does the VV 10th Anniversary pedal have any other "protection" for the Max1044?

It has some careful considerations of the incoming RC time constants on the +9V line, and one subtle layout trick that I came up with that I think protects it in some very esoteric cases. No magic circuits, no.

FWIW, I have sold the 1 Sopt at retail for several years and have never heard of one frying a pedal. We did see a very high failure rate with the 1 Spot itself when we first started selling them but they have been extremely reliable for the last few years.

That corresponds to the first vendor who supplied us 1Spots, I think. It was before my time, but I do know that there was one early manufacturer who had poor consistency. VS changed to a much more capable manufacturer a bit before I started.

But if its static that's killing the chip, than the 15V or so that the 7660 can stand is still not very much.

No, it's not. However, it may be related to the open pin 6 and 7 I mentioned. The app note says to leave them open, even, and that makes me crazy from long experience with CMOS.

I would suspect that the 1Spot is sometimes producing a little over 9.5V temporarily, and since that voltage is already close to the max of the MAX ... Has anyone fried a MAX with another power supply than the 1Spot?

I suspected that to - and that's part of why I did the elaborate testing. It may be there, given that there are many thousands of them in the field, but I can't find it by beating on a number of 1Spots trying to make it happen.

To answer your question, yes; I've fried a MAX1044 with a power supply other than a 1Spot. You see, the 1Spot is not the only switching power adapter I subjected to the tender mercies of my test bed. There is another widely advertised switching 9V power supply for effects, and two other 9V supplies that have been suggested at places I found on line for powering effects. So I ordered "victims" and beat on them all separately but equally. I can't say much about the others except that from what I saw, I would not run my pedals on them.

So far, my Boss PSA has not given me problems, but I have only played my MAX-pedals at home, and not under more "extreme" conditions.

The Boss PSA is a standard low frequency transformer->rectifier->filter->linear regulator. It's almost impossible for it to give you problems **unless** the linear regulator shorts. When that happens, the output voltage will go to the internal voltage of about 12.5V, and your MAX will die, unless the short is to ground, in which case the internal fuse will (probably) open. Not the same animal.

[Deric]

It's important to note that I have no idea what has caused 3 Max1044's to go bad in my PT-80. It just so happens that I use a 1 Spot - I do not know if that has anything to do with it. This also happened over a long period of time - about 2 years. Although it's been a while, I MAY have been using a PSA-120 when the first one went bad.

I stumbled across the 7660S at DIYStomps and figured it was worth a try. I made no changes to the circuit - simply un-plugged the 1044 and plugged in the 7660S. It's only been a couple of weeks so..........

Tonight I swapped the 7660S and 1044 back and forth and listened for noise issues. I could hear no difference in noise or tone/performance with either one. Out of curiosity, I also measured the output of a PSA-120 and it was higher than the 1 Spot ( 9.56vdc vs. 9.4vdc ).

RG-

and one subtle layout trick that I came up with that I think protects it in some very esoteric cases

Care to elaborate?

[R.G.]

It's important to note that I have no idea what has caused 3 Max1044's to go bad in my PT-80. It just so happens that I use a 1 Spot - I do not know if that has anything to do with it. This also happened over a long period of time - about 2 years. Although it's been a while, I MAY have been using a PSA-120 when the first one went bad.

I stumbled across the 7660S at DIYStomps and figured it was worth a try. I made no changes to the circuit - simply un-plugged the 1044 and plugged in the 7660S. It's only been a couple of weeks so..........

Tonight I swapped the 7660S and 1044 back and forth and listened for noise issues. I could hear no difference in noise or tone/performance with either one. Out of curiosity, I also measured the output of a PSA-120 and it was higher than the 1 Spot ( 9.56vdc vs. 9.4vdc ).

Yeah, the PSA uses a 7809 regulator with a diode from it's "ground" terminal to the actual ground. This normally gives you about 0.45 to 0.5V, that being typical for a rectifier silicon junction at the low currents in that chip's ground path; however, the 7809 has a +/-5% tolerance, too, although most of them are within 2%. It's easy enough to get there.

There's a tolerance on every electronic value. The basic reference voltage in the 1Spot is about +/- 1.5%, but it has to be modified to 9V with resistors, so the resistor tolerances add in there too. I've seen 9.5V 1Spots on the high end.

Care to elaborate?

Sure. You just have to make sure that any trace which includes connections to both the MAX1044 power line and the flux capacitor is no wider than 1.44 gigahertz, for obvious reasons. It's pretty simple when you think about it.

[Deric]

is no wider than 1.44 gigahertz

That explains everything!

[WormBoy]

Change your MAX1044's to LT1054's - they can put out 10 times the current, and have a max of 15V supply!

Following your advice (thanks! ) , I bought two 1054's and just put one in my PT80 delay. I bent pin 1 out of the way as this has a different function than in the 1044's, and is probably not supposed to be attached to 9V (from glancing through the datasheet). So far it works like a charm. In this circuit (GGG board) it works like a voltage doubler and puts out some 16V which is enough for the 12V regulator that follows.

[Danch]

I have a question if someone could answer, soundwise is it the same if I power the pedal with for example +18/0V instead of +/-9V, headroom is the same 18V and there is no differences or?