Grandmeister 36 - does not appear to send midi output from amp

No, forgive me for my pedantic nature, you have clearly sorted that out for me now. I am confident that what we think we are seeing is just as you say it is.

Can't remember if you have said this but, have you performed a factory reset on the GM36? This would delete all of your stored presets on the amp of course and change them back to the factory set so I realise it may not be possible at this moment if they are a custom set you need.

Hi Fredo,

I started the app with Midi Monitor showing Input and Output. The red output panel shows data.

I moved knobs on the amplifier: Channel (Clean to Crunch), changed reverb setting, etc... No data is displayed in the green input panel.

I tested the Roland UM-ONE USB/Midi interface using MIDI-OX. The device seems to be working correctly by sending and receiving data.

I suspect that something has gone wrong with amplifier midi OUT port. I was hoping that there might be some way to validate if my suspicion is correct and that maybe there is some simple solution.

Hello bordonbert,

"pedantic" - I had to get my dictionary out for that one :-). You must have a vast lexicon. You just added one more word to mine. I also am a bit pedantic - my wife calls it OCD.

I have not factory reset the amplifier. I was holding off on that in case there was some other solution.

I can certainly do the reset. I have settings stored for specific songs but I don't know the preset number without the help of the remote app. I do have a backup of the amplifier in the form of a ".gm36" file on my PC. If I ever get this working again then I can reload those settings and only lose some recent changes.

Do you think it is time to pull the trigger and factory reset?

I don't think it would hurt to try it. I will say, Fredo is the expert in this field by a country mile. That app is his own creation, that's how good he is in this field. Anything he adds to this conversation should really be listened to hard. As long as you know you have your presets in the amp stored somewhere then I don't think it would hurt to try that, it can sort out the strangest of symptoms.

There is a way to check out whether the amp MIDI output is working but it would involve putting an oscilloscope on the output to monitor when data is sent. You can't just measure voltages on an output pin of a MIDI system to ground as it is a balanced signal. That means it has no connection or reference to ground. The two output pins carry a +ve and a -ve version of the same data signal and they rely on the input of the "listening" device to set the level that they vary around. Effectively they "float" but for signal purposes they reference each other.

I wouldn't suppose you would have an oscilloscope in your toybox would you?

I was aware that Fredo is the author of the GM remote apps.  I was very happy when he joined in the thread.  He must be a midi and programming genius.  If we get one of the H&K engineers to join in we will have the world authorities :-)

Unfortunately, I do not have an oscilloscope or other fancy sniffers.  I used to have access and used such equipment when I was working but I have been retired for some time now.

I will wait until tomorrow morning to see if any other suggestions come in before performing the factory reset.  It is about 7:00pm in my time zone now.  That brings me to my next question.  It is none of my business but I am wondering if you ever sleep.  You have posted messages at all hours.  It must be going on 1:00am in Southern England now.  

Thanks again for your help and sharing your knowledge.

I performed a factory reset on the amplifier.  The reset was successful but it did not resolve the problem. I will be taking the amp back to church tomorrow.  I love the tone of this amplifier.  I will miss the remote app.

Thanks to bordonbert and Fredo for sharing your time and knowledge in an attempt to resolve the issue.

Hey bordonbert,

You wrote that the midi-out port could be tested with an oscilloscope.  I moved from Michigan to Tennessee about one year ago.  It turns out that a member of the church I attend here is an electrical engineer that builds amplifiers in his spare time.  He will have an oscilloscope.  Where might I find enough specs to give him so that he can diagnose the midi-out port?

Thanks.

Don't worry about specs, you don't need to be at all rigorous about any of this.  The signal is on pins 4 & 5 but the DIN socket pins are not numbered sequentially.  The socket will have 5 pins arranged equally spaced in a semicircle.  The ones you want are the second ones in from each end, that's the two on the diagonals.  I've added a pinout diagram for the plug and socket. Use the right hand female pic as a reference for the socket on the back of the GM36. You should expect signal voltages of around +5V.

There are two basic ways to easily check the output function.  Just stick a resistor in pins 4 & 5 and put your scope probe across that.  (This assumes the scope input is floating so you will not be grounding anything by doing that.)  Anything from about 560ohms up to a few kilohms will do.  You should be able to see the pulse chain for an instant when any change is made. A simpler approach which you may be able to use is to put a LED in the pins of the socket with its anode (+ve) terminal in pin 4 and its cathode (-ve) terminal to pin 5.  When you turn any of the controls you should see the LED flicker.  That mimics the action of the LED which is in most MIDI input circuitry.

Using the scope, when you make a change to the amp settings you should simply see a burst of pulses pushed out.  It will be short and sharp but you can see it.  Rotating the knob back and forth constantly should give you a stream of pulses which will make seeing it much easier.  If the scope is a storage type then you will be able to capture it.  Decoding what the pulses mean is not important as we are just trying to ascertain whether the amp is sending data out or not.  As long as you can detect pulses each time don't worry about what they are unless you are like me, "curious" is putting it kindly, "obsessive" is being a bit more realistic.

If you really want to get to grips with things and you have that storage facility to capture the pulse stream then this video gives much more info than you really need:  Reading MIDI with a Scope.  He actually uses extra electronic circuitry as an interface but don't worry, you don't need that to see what is going on.  Incidentally, it doesn't make any difference to what we have said but for the record, I described MIDI as a "balanced" signal system.  It isn't!  I was in the wrong thinking mode there in terms of what I used to work with.  It is actually a 5mA current loop.  Not a big deal but we had better leave things accurate and - when I'm wrong I'm wrong.

This should at least tell us whether the amp is functioning on some level and sending MIDI data out to the interface and app.  If you have a storage scope we could even decode one of the messages to see if it is accurate but that shouldn't really be necessary at all.

Got a LED handy?

I had a box full of electrical components (resistor, capacitor, led,. . .) left over from my college days. I got rid of that, and a lot of other things, while downsizing for our move to Tennessee a year ago. The closest store for such things is at least an hour from where we live now.

Josh is the electrical engineer that I wrote about. I hope to be in contact with him sometime in the next several days.

Thank you for the detailed information. I will let you know what Josh finds. I am looking forward to that.

Conclusion.

I finally met with Josh (electrical engineer) yesterday.  We confirmed that the midi out port is no longer functioning.

Thanks bordonbert and fredo for your support.

Dave

Damn! That is bad news and a very unusual situation to have Dps2. MIDI hardware is normally very robust and problem free with its current loop design. Sorry to hear you couldn't get it up and running without having to delve inside the amp. What is the way forward for you then? Is Josh going to be able to dig into it and repair it? The normal solution would be to replace the whole of the Rearboard but that isn't necessarily the cheapest way depending on who is fixing it.

The MIDI components are all on the Rearboard and are easily accessible. The only thing which could really be expected to pop is the output chip, (we know the input is working). The difficult thing is that the components are surface mount devices and awkward to work with without fine tools. If Josh wanted to see if he could take a look and maybe replace the output chip, (SN74LVC1G08, not terribly expensive at around £5 for 10 on ebay but very small), I could help him with info as to what to look for and where to find it inside.

Let me suggest a couple of other "not really likely" things which he could do more easily before starting to replace anything.

Measure the voltage on Pin4 of the output socket with nothing plugged into it.  It should be a solid 5V as it only connects direct to the +5V power line inside through a 220R resistor.  That is the second from the left on the left diagonal as you look at the socket.  If you are looking for the MIDI signal on the bare pins of the socket, you won't see voltage signals coming out of either pin with a meter or a scope, MIDI is a current loop setup and you can only see signal voltage between the output and a higher voltage line.  That is why pin 5 feeds the signal from the output device through a 220R resistor to pin 5, across the input of the listening device to pin 4, through another 220R resistor to the +5V line.  You must complete that loop for any signals to appear!

MIDI is a strange system compared to simple signal lines and you need to set it up right just to see the signals being outputted.  You need to close up the loop by connecting up pins 4 and 5 with something which will respond to the current which is the true output.  You can mimic a listening device on the line by inserting a resistor of about 470ohms or the LED I mentioned across pins 4 & 5, (LED is the best way but get it the right way around).  Then look for the LED flickering or measure the voltage directly across the resistor, not from either terminal to ground.

If the problem goes deeper than just the output, the MIDI area of the Rearboard is fed the MIDI output signal from the Microcontrollerboard via a short 16 pin interboard cable.  The MIDI Out signal is on pin 12.  If you are inside you could try breaking down and remaking both ends of that.  Flex the cable a little at the plugs to check for any faulty connections between the wires and the pins.  If this is a problem, flexing it usually shows it up by breaking any corrosion on the connection at least temporarily.  Gently take the plugs out of the PCB sockets at both ends and reinsert a couple of times to wipe and clean the contacts.  These are high quality aviation standard connectors so they are usually trouble free but you never know and, if you are inside the amp already, it costs nothing to check this.

What equipment does Josh have to hand to test things?  Are you looking to go inside, (not the easiest thing to do with an amp of this high tech miniaturised type).  On the other hand, are you thinking of putting the amp in for repair to a shop?  They would have access to service documentation and replacement boards in a way that you and Josh won't.

Wow!  I did not expect the lengthy and very detailed explanation.  I am very grateful for all your support.

We did take off the end cap and bottom plate.  We were hoping to visually inspect the midi socket for faulty connections.  The PCB blocked view of the socket.  Without the ability to perform diagnostics, we just closed it back up.  The amps that Josh builds are 60's style hand wired tube amps.  Neither of us wanted to risk causing additional damage to the amp by digging further into it.

The Grandmeister is otherwise functional.  I have simply lost the benefit of using an app to manage the settings.  I plan to continue using the amp as-is until something else quits.

Thanks again.

I've got to be honest Dps2, I'm trying to be a bit diplomatic here in the advice that I'm giving so as not to discourage you.  I know you are dispirited because there is nothing to see but that is very much what I expected. You won't see a cracked joint at the socket without high magnification.

I get the point that Josh is used to working with classical type circuits and components, SMD devices as we have in the H&K kit are a nightmare to those used to through hole large components and rat's nest wiring.  They are not lacking in skill, they just have a different skillset.  Despite Josh saying there is a fault, I still don't believe that a hardware fault is the answer.

An aside to preach a little, humour me a tad please. Despite what some old school techs and all internet scuttlebutt tell you, the modern approach with PCBs and automatically soldered SMD devices is way more reliable and is so much better in terms of performance repeatability it is a joke.  It's just that most of the old guard (and I'm an old man in that group too) can't get to grips with fixing it as easily when it goes wrong and you can't use a poker for an iron with an inch between joints.  Hence the insistence "it is c**p" and the logical extension to rub that in with the gullible, "it sounds s**t." Well on the bench you can clearly show that the idea of a "boring sterile modern sound" is a total myth.  There is no reason for an older valver to sound different to a modern one like the GM36, it's all down to the design aims and implementation.  That's what the H&K amps are all about.  They use valves for their qualities with more modern implementations to give great versatility - and reliability!  The number of modern electronic devices that go wrong is much much smaller than the number of old tech ones when calculated as a proportion of how many are sold.  Wghen they do the repair is much more difficult, hence the idea of "just replace the board". Old style - easier repair.  New style - much less repair to perform.  You pays your money and you makes your choice.

Now back to your amp and fixing it. Have you tried the LED test I suggested?  It takes one cheap component and it would let you know immediately whether anything was coming out of the MIDI socket or not in a really simple and obvious visual way.  You see I know the circuitry of a MIDI output circuit well, (much more than I know how to use it, we have others around who know a lot more than I do on that score. )  In terms of damaging it there is very little you could do to it other than hooking it up to a high voltage which can cause it a problem!  Accidentally grounding either of the pins 4 or 5 or shorting them together will do nothing, protection against that scenario is built in.  There is a 220R(ish) resistor in each of the lines feeding the socket pins and this limits any fault currents to safe levels well below the continuous rating of the chip which drives them.  MIDI was designed for a musical market where everyone has days when their fingers are like hotdogs, some are in that state as a matter of course.  Clumsiness was built into the design so it is virtually bombproof.

My gut feeling, and I could be wrong of course, is still that there is nothing wrong hardware-wise with the MIDI output drive circuitry. The flickering LED test would clearly indicate whether there is any data coming out of the socket. It may be that this is being scrambled in some way so your interface doesn't recognise it as valid, but I find it hard to believe it has stopped pumping out data completely through a hardware failure.  I can be wrong in that of course but it makes sense to apply that simple test to confirm what you currently just believe.  Remember, detecting signal voltages on a MIDI output is not the way it should be tested. They work with currents and they aren't detected in the same way.

I was hoping that Josh would have an LED and resistor that we could use to test the midi out as you described.  He has resistors, of course, but unfortunately no LED.

I did my homework to understand the method you described.

This video shows how to make a tester and is very much what you described:
https://www.youtube.com/watch?v=49t9outj_jY
-  Solder resistor to either end of LED
-  Solder cathode end of LED to pin 5
-  Solder anode end of LED to pin 4

This page has a calculator which tells me that I would need a 150 ohm resistor for a standard 3-5 mm LED for the 5 volts expected from the midi out:
https://kitronik.co.uk/blogs/resources/led-resistor-value-calculator

I could buy a package of LEDs on Amazon for about $5 USD.  But, if I the test shows that I am getting no output or scrambled output then I am still in the position.  The midi-ox test that I did and response from Fredo ("Dps2, the screenshot shows that the app is not receiving any message from the amp...") seems to indicate that data is not being sent or is scrambled.

I might be wrong (would not be the first time) but it looks like all roads lead to the repair shop.  The closest authorized H&K dealer is at least 200 miles (320 km) away.

I am content to use only a few amp settings for the music that we do at church.  I can keep track of those and call them up with the FSM.

I am totally appreciative of your all the help that you have been and your persistence to help me solve this problem.

Full marks and credits for doing your homework but in this case it's actually much easier than what you describe. The resistor you describe calculating and adding to the LED is absolutely unnecessary. The MIDI output circuitry already has resistors in it to do that job. Likewise you don't need to open the amp up. Just poke the legs of the LED into pins 4 and 5 of the socket.

This is what Fredo eventually wrote: "Dps2, the screenshot shows that the app is not receiving any message from the amp. The amp should respond to the command that is sent every second. Maybe it does, but the application is not receiving it... If you don't see anything in the input (green) window while rotating knobs on the amp, it means that the MIDI connection between the amp and the tablet is not working for some reason :-/"

Have a look at the bold bits. The MIDI monitor inside the app cannot say that the amp is not sending data, only that it isn't getting as far as the app. There may well be a problem with the interface selection or its setup or even the interface circuitry could be damaged. The app cannot tell that, it can only say "well it isn't reaching me". If the amp sends out the data and the interface can't decode it and pass it on to the tablet and ultimately the app then the app cannot respond in any way or show that there is any data in the first place. Remember this isn't just the app talking directly to the amp like it does to the FSM432. Think of the interface like the exchange in a phone conversation. When you are on the phone to someone and something happens at the exchange what do you do? You say "Hello, hello, are you still there?" It looks to you as though the caller has ceased to exist but they will be doing the same thing to you at the other end of the line and they will think you have ceased to exist too. There is nothing wrong with either your phone or theirs, both are still sending out signals but the exchange cannot pass them on.

I think your final position regarding being able to use the amp as it is now is a good one and I'll quiet down with the ideas. If you can get to a tech and you are not averse to paying him to look at the amp then I think that is a good thing given the circumstances. I would suggest you take your Roland UM-1 interface with you too so he can test the amp using that as well as something he has which he knows is working. If there should be a problem with the interface then he would spot that too - all in the same price! Do keep us up to date with how you get on and let us know what you find, it is an intriguing problem. You never know when someone else can benefit from your own experience. I'm sorry if I seemed to bash on about this, I am known as a pedantic old ba.....rnacle but it is to your advantage to absolutely rule out every eventuality we can before you take that long journey and hand over your much needed bucks for a fix you may not need.

Good luck.

Okay bordonbert.  I will go the last mile and come up with an LED.  It will take me a few days and I will let you know.

Insert cathode end of LED to pin 5 and anode end of LED to pin 4.  Right?

Thanks again my pedantic dear friend   

Hahaha!   The Pedant!!!  Now you're seeing into the soul of the man now aren't you?  It's simple, cheap, and there is just a chance it might save you a long expensive round trip.  And is it going to take you more time to find an LED and try it than the delay until you are able to travel to your tech?

Yes, anode to pin 4, cathode to pin 5.  If it is a new LED then the legs will be slightly different lengths.  The anode is always the longer leg.  Don't worry if you get it the wrong way round, it's a diode so it will just block any current through the line.  The low +5V line on pin 4 means it is not taken beyond its reverse voltage.  If nothing happens try it the reversed way round just in case you have made a mistake, as I said it will damage nothing.  The 220R resistors inside will protect the LED even if it were under threat.  And don't worry if you find it stays on all the time.  That might just be a sign of how the driver is sitting inside.  What we are looking for is flickering of the LED when you turn a knob.  The lack of that is the sign that something is wrong in pumping MIDI data out and we are now testing direct at the amp with nothing between to confuse things.

As you rightly pointed out, you can use the amp in its basic form in the meantime, though I agree, it's highly frustrating to have to do without the capability you know you really want.  My wife always tells me "Patience is a virtue, possess it if you can.  Seldom found in woman - never found in man".  She can be brainless at times but, if I can paraphrase, "Out of the mouths of babes and sucklings come forth things you just don't want to have pointed out by your wife".

Hi bordonbert,
I ordered 5mm LEDs from an online store and received them earlier this week.  I tested a few of the LEDs across 3 volts to make sure they were good and then took those to the church yesterday to test the midi out across pin 4 and 5 as we discussed.  Sadly, there was no activity shown on the LED as I adjusted amp settings.  I have to conclude that the midi out has stopped working for some reason.

Thanks once again for your help and persistence and encouragement in diagnosing this problem.

Great stuff.  Glad you took the pro route of ruling everything out that you could first.  But a bit of advice for the future...

"I tested a few of the LEDs across 3 volts to make sure they were good...".  Be advised you should always add a current limit resistor to the LED!  LEDs are not bulbs.  Bulbs are designed to want to see a particular voltage and their resistance is designed to draw a specific amount of current from the supply at that voltage to allow correct function of the bulb.  You cannot just put LEDs across any voltage source and hope they will work as they should.  They are not suited to responding to voltages, they need you to feed them a defined and limited current and the voltage across them then settles at the correct level for that voltage.  In their active region small changes in voltage mean huge changes in current so the LED is uncontrollable in that manner.  However, small changes in current mean tiny changes in voltage which means good control, so current is the parameter with which we control them.  Even when you seem to be controlling them with a voltage, as in your case, they are still really responding to their current.  (See my comments about drycell batteries lower down.)

If that 3V source you used is low enough impedance, (effectively if it has nothing inside which would limit its current to safe values), then the LED would be forced to sit at 3V and this is well above what most of them are intended to take, the red ones at least.  At 3V they would draw very high currents, high enough to damage or destroy them.  You need to have a resistor in series with them to soak up the excess voltage above what the LED would be happy with.  The resistor then limits the current through both of them to a safe value for the LED.  Now, in real life, if the 3V source is a simple drycell battery or two, these have enough internal resistance inside for them to act as that current limit resistor.  Hopefully that would be your case and the LEDs would not have been harmed.  I worked for years with an internationally very highly respected seismic survey engineer who simply could not grasp this point and kept blowing LEDs on site by continuously sticking them onto voltage supplies without any resistance.  He was always puzzled when they blew, and I never did get him to see the point.  "But it lights up, it's just a bulb!"  

That said, that was just a heads up for the future.  I think you have acted above and beyond the call of what most guitarists would have bothered with and I take my hat off to you.  This is MIDI and not just the audio action of your amplifier so it would be beyond the thinking of most musicians to investigate it, (understandably).  Well done for persevering, I will add "with me" to that .  The only answer now is just as you say, to get a tech to look into the guts of the amp and investigate.  It may be something as simple as a bad connection between boards which can be fixed by simply disconnecting and cleaning contacts, (that's unlikely), or a cracked joint on the PCB in the MIDI Output area, (not common but possible), which will only require a resolder.  However, if the problem turns out to be damaged components in the MIDI Output section then you will need to find a tech comfortable with replacing SMDs (surface mount devices) on the PCB.  Your only other alternative would be to order a whole new Rearboard PCB from H&K!  I don't even like to think about the cost of that.  On the other hand, for the price of a cheap chip, (£4.99 for 10 here in the UK so I would guess cheaper in US), it shouldn't break the bank.  This is not something that every guitar amp tech can or will want to do.  Many are only comfortable with working on large style components in point to point "rat's nest" wiring of old style amps.  (They aren't comfortable with it, hence the myth that "PCBs make amps sound c**p".    )

I can maybe help a little with your dealing with the tech.  I'll PM you with some details, check your mailbox in a short while.  But for now, well done on getting as far as you have.

Good advice about the voltage across the LED.  The set of LEDs that I bought came with 2.0-2.2V and 3.0-3.2V variety.  I put 3V across one of the red ones (2V) without a resistor just to see what would happen.   I could not stop myself from trying that :-).  It burned out quickly.  I used the 3V LED to test the amp and then tested the LED once again when I got home to make sure that it was still good.