Standard midi driver circuit will not drive wireless midi transmitter - need circuit design help.

Thread Starter

Bellows

Joined Jan 20, 2018
13
I have built a simple midi out circuit feeding midi data through two inverters intended to drive a commercially available midi transmitter. The circuit works correctly when driving a standard midi input to a music device, but generates gibberish when connected to the midi in of a wireless midi transmitter. I need help to diagnose what is going wrong and modify the driver so it will correctly drive the transmitter.
 

dendad

Joined Feb 20, 2016
3,537
One of the first things you need to do is post the FULL circuit of your device. ALL the connections. Assume nothing as often the parts assumed are the problem.
Post any info on your "commercially available midi transmitter" too.
Without more info we cannot help you.
 

Thread Starter

Bellows

Joined Jan 20, 2018
13
One of the first things you need to do is post the FULL circuit of your device. ALL the connections. Assume nothing as often the parts assumed are the problem.
Post any info on your "commercially available midi transmitter" too.
Without more info we cannot help you.
Yes, I understand. Had planned to post additional details. After further observation, I have determined that there is an anomaly I need to "reverse engineer" and clarify inside the complex proprietary midi controller in order to fully characterize what's happening. It may not be a midi driver issue. I will elaborate when I have completed this. In the meantime, the description of the problem remains incomplete.

Thank you for your reply.
 

absf

Joined Dec 29, 2010
1,943
Here is a typical MiDi interface from a Music keyboard. Is the interface you design similar to this?
midi in music kbd.PNG
Some models have a MiDi through connector.

Allen
 

Thread Starter

Bellows

Joined Jan 20, 2018
13
Complete driver schematic design schematic is attached.

Here are the observed symptoms of the problem:

When this driver output is hardwired to the midi module, it works correctly. The connecting adapter cable converts from midi connector at the output of the driver to a db9 at the input to the midi module, using only three pins of the db9. In this case the logic low levels are 1.0 volt or lower.

When the driver output drives a Midi Jet Pro https://www.midijet.com wireless system, with the output of the receiver connected to the midi module using the very same adapter cable (midi to db9), only two of the midi channels are correctly transmitted to and received by the midi module. The third channel, midi encoded for the keyboard of the accordion controller, is corrupted or degraded. The logic low level is above 1.0 volt.

However, when the output of the midi jet receiver is connected to a typical midi keyboard, etc., all the midi data is correctly received and interpreted.

An identical circuit - albeit with different brand components, i.e. hex inverter and voltage regulator, is used in an almost identical controller/module combination, together with the midi jet pro, and works well.

I also have a schematic of the midi received at the input to the midi module which i can upload if needed.
 

Attachments

dendad

Joined Feb 20, 2016
3,537
Just a couple of points...
In your top circuit, is there a Gnd on the inverter supply as you show both inverters hooked to power?
In the bottom circuit, all unused inverters need to have their inputs connected together Gnd, Vcc or another inverter output. Never floating.
Leaving them floating could be generation random noise signals and increase your power consumption bit.
 

Thread Starter

Bellows

Joined Jan 20, 2018
13
Just a couple of points...
In your top circuit, is there a Gnd on the inverter supply as you show both inverters hooked to power?
In the bottom circuit, all unused inverters need to have their inputs connected together Gnd, Vcc or another inverter output. Never floating.
Leaving them floating could be generation random noise signals and increase your power consumption bit.
Thank you.
Yes, the inverters are supplied with regulated VCC and with VSS.
Thanks for the tip to ground the unused inverters. These are not grounded on the reference circuit i copied buti can certainly ground them. But, does that in any way address the problem i am experiencing?
 

dendad

Joined Feb 20, 2016
3,537
Thank you.
Yes, the inverters are supplied with regulated VCC and with VSS.
Thanks for the tip to ground the unused inverters. These are not grounded on the reference circuit i copied buti can certainly ground them. But, does that in any way address the problem i am experiencing?
It may not, but then again, it may be the working circuit is injecting noise the proto did not do. Either way, it is a trap a lot of us have all fallen into, leaving inputs unconnected, and it can cause some strange behaviors. I often just connect the unused input to an output next to it. That does not matter if the unused parts are running like that, with their output not connected. Just as long as their inputs are correct.
 

Alec_t

Joined Sep 17, 2013
11,324
1) Can you confirm whether the dodgy driver uses a 7404, 74S04 or 74LS04 hex inverter?
2) The datasheet for the NTE977 regulator advocates a minimum load capacitance of 10nF to limit the high frequency noise bandwidth.
 

bertus

Joined Apr 5, 2008
20,540
Hello,

Looking at the handdrawn schematic, an open collector inverter should be used.
Wich inverter chip are you using?

Bertus
 

dendad

Joined Feb 20, 2016
3,537
And so you have bypassing caps in your circuit? I'd put a 100nF ceramic as close as you can, like directly from pin 7 to pin 14 of the 7404, and a 10uF tantalum cap on the 5V supply too.
 

dendad

Joined Feb 20, 2016
3,537
Looking at the handdrawn schematic, an open collector inverter should be used.
I actually missed that! My excuse is it is early morning here, and not enough coffee ;)
Yes, it should be open collector, like 7417.
 
Last edited:

Thread Starter

Bellows

Joined Jan 20, 2018
13
Ok - don't find a datasheet for my NTE7404 inverter which i bought at Fry's.
looks like i need to go shopping for...
a) 10uF tantalum cap (don't know how big this is physically)
b) a 7417 (this is open collector?)
c) small 100nF ceramic
The unit is at a friend's house remote from my present location. Preparing for a debugging session next Thursday. Any other help would be appreciated to prepare for that with relevant components and have all i need in hand. (As a former MOS IC memory designer, don't have experience with TTL and systems components.)

Also, would like to invest in an oscilloscope for projects like this. Recommendations appreciated.
 

dendad

Joined Feb 20, 2016
3,537
Also, would like to invest in an oscilloscope for projects like this. Recommendations appreciated.
There are a lot of +100Mhz scopes around now for about $300.
Rigol have a pretty good name and so does Hantek I think. This looks pretty good to me..
https://www.ebay.com.au/itm/Hantek-DSO5202P-200MHz-2-CH-1GSa-s-7-TFT-LCD-Digital-Storage-Oscilloscope/282764929924?hash=item41d61a6b84:g:xKYAAOSwFb5aKpgu
I have a number of older scopes. They are not referred to as "Truth Meters" for nothing. A good scope is a well worth while investment.
 

Thread Starter

Bellows

Joined Jan 20, 2018
13
Yes, the 7417 is open collector BUT it is a buffer, not inverter, so you just use one, (not 2 in line like your first circuit).
http://www.ti.com/lit/ds/symlink/sn7407.pdf
As bertus mentioned, if you want inverters, use 7416.
http://www.ti.com/lit/ds/symlink/sn7406.pdf
I addressed everyone's advice. I'm at an impasse.

To summarize, it is a Version 3 Millennium accordion i am attempting to modify to become wireless midi. My newly built driver works correctly in another similar (later version #4) of the same brand accordion i am attempting to make wireless. When I substitute my newly built driver into the later version 4 accordion, which was already wireless from the vendor, it plays just like with the vendor's original driver. The midijet is in place of the cable and all parts work as expected.

With my new driver in the earlier Version 3, only the left hand plays correctly when i use the midijet wireless system But both sides of the accordion work correctly when i hard wire the driver to the sound module, replacing the wireless system with the cable, i.e. very the same adapter, i.e. midi to db9.

When I look at the signal levels and general character with my newly acquired scope, everything looks similar. No obvious difference when i look at the output of the wireless system/input to the module vs. the end of the hard wire cable/input to the module.

Prior to this latest test, midijet support people said something about suppressing the clock data from being transmitted. I have no access to the micro code, so don't know of a simple way to do that, short of building a logic circuit that would subtract out the clock. That seems a bit arcane. But the solution must have something to do with the amount of data that the wireless system can handle. And, just maybe, the later version 4 accordion had a software (or wiring?) change to eliminate the clock....or some such thing.

I expect at this point the solution might need to be implemented in software. Any body have any ideas or know of a resource to help me with this?
 
Top