Yamaha DX7 memory backup circuit

[[email protected]]

I'm restoring a mistreated DX7 to working condition.

Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png

The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board).

/IC is an "initial clear" signal that comes from the power supply.
From staring at the schematic for a few hours, it looks like during
power-on, the RAM is kept powered by the battery, and then when the
power has stabilized, Tr1 switches the 5V rail to the RAM (stopping
current from the battery via D4). Also, during power-on, it looks like
Tr3 prevents the /CE2 line from asserting (there are four RAM IC's,
and /CE2 doesn't go anywhere else on the schematic.) I guess this
prevents garbage from getting written if /WE wobbles during power-on?

So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter? Is current an issue? I wish I had
another DX7 to examine but I don't. I can't find a datasheet for the
M5M5118P so I don't know what kind of Vdd it wants.

Thanks as always..

(BTW the schematic came from this page:
http://www.maths.abdn.ac.uk/~bensondj/html/dx7.html#diagnostics )

 

[Franc Zabkar]

I'm restoring a mistreated DX7 to working condition.

Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png

The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board).

/IC is an "initial clear" signal that comes from the power supply.
From staring at the schematic for a few hours, it looks like during
power-on, the RAM is kept powered by the battery, and then when the
power has stabilized, Tr1 switches the 5V rail to the RAM (stopping
current from the battery via D4). Also, during power-on, it looks like
Tr3 prevents the /CE2 line from asserting (there are four RAM IC's,
and /CE2 doesn't go anywhere else on the schematic.) I guess this
prevents garbage from getting written if /WE wobbles during power-on?

So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter? Is current an issue? I wish I had
another DX7 to examine but I don't. I can't find a datasheet for the
M5M5118P so I don't know what kind of Vdd it wants.

http://www.datasheetarchive.com/pdf-datasheets/Datasheets-110/DSAP002745.pdf

The operating voltage is 5V +/- 0.5V.

"Data can be held with 2V supply voltage".

It seems to me that a general purpose silicon diode should suffice.
The current draw of each RAM in standby mode is only 15uA.

Thanks as always..

(BTW the schematic came from this page:
http://www.maths.abdn.ac.uk/~bensondj/html/dx7.html#diagnostics )

- Franc Zabkar

 

[Bob Larter]

I'm restoring a mistreated DX7 to working condition.

Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png

The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board). [...]
So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter?

Some random, small Schottky (for the low Vf) diode would probably do the
trick. I've seen germanium diodes used in similar circuits, but I
wouldn't recommend one with a lithium cell, as it will leak a small
current into the cell, which they don't like.

Is current an issue?

No. The standby current on those low power SRAM chips is measured in
microamps.

I wish I had
another DX7 to examine but I don't. I can't find a datasheet for the
M5M5118P so I don't know what kind of Vdd it wants.

5V when active. In standby mode they'll retain data down to a couple of
volts.

HTH!

 

[Jamie]

I'm restoring a mistreated DX7 to working condition.

Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png

The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board).

/IC is an "initial clear" signal that comes from the power supply.
From staring at the schematic for a few hours, it looks like during
power-on, the RAM is kept powered by the battery, and then when the
power has stabilized, Tr1 switches the 5V rail to the RAM (stopping
current from the battery via D4). Also, during power-on, it looks like
Tr3 prevents the /CE2 line from asserting (there are four RAM IC's,
and /CE2 doesn't go anywhere else on the schematic.) I guess this
prevents garbage from getting written if /WE wobbles during power-on?

So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter? Is current an issue? I wish I had
another DX7 to examine but I don't. I can't find a datasheet for the
M5M5118P so I don't know what kind of Vdd it wants.

Thanks as always..

(BTW the schematic came from this page:
http://www.maths.abdn.ac.uk/~bensondj/html/dx7.html#diagnostics )

Use a Schottky diode, they have good handling current and low forward
voltage.
Also, you may want to change the electrolytic cap just after the diode
to prevent drainage. normally a high Q cap for this is a good idea.

http://webpages.charter.net/jamie_5"

 

[Franc Zabkar]

I'm restoring a mistreated DX7 to working condition.

Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png

The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board). [...]
So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter?

Some random, small Schottky (for the low Vf) diode would probably do the
trick. I've seen germanium diodes used in similar circuits, but I
wouldn't recommend one with a lithium cell, as it will leak a small
current into the cell, which they don't like.

I've always thought that Schottky diodes were fine for the OP's stated
purpose, but after you raised the point about reverse leakage
currents, ISTR that Schottky diodes aren't too good in this regard.

- Franc Zabkar

 

[Fox]

I'm restoring a mistreated DX7 to working condition.
Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png
The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board). [...]
So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter?

Some random, small Schottky (for the low Vf) diode would probably do the
trick. I've seen germanium diodes used in similar circuits, but I
wouldn't recommend one with a lithium cell, as it will leak a small
current into the cell, which they don't like.

I've always thought that Schottky diodes were fine for the OP's stated
purpose, but after you raised the point about reverse leakage
currents, ISTR that Schottky diodes aren't too good in this regard.

- Franc Zabkar

Interesting.. I hadn't considered the reverse current aspect. I was
only worried about the voltage drop. Since you got me on the right
track I found this:

http://www.panasonic.com/industrial/battery/oem/images/pdf/Panasonic_Lithium_DesignMemoryBU.pdf

It gives a rough idea of the maximum allowable reverse current
depending on the application.

 

[Franc Zabkar]

[email protected] wrote:
I'm restoring a mistreated DX7 to working condition.

Re this schematic:
http://www.bobdbob.com/~deneb/ftp/dxmemory.png

The battery is a lithium coin cell, like a CR2032. D4 on my board has
been replaced with a gigantic power rectifier. I'd like to re-replace
it with a part more appropriate for the application (or at least that
fits through the holes on the circuit board).
[...]
So what I'm wondering is, what kind of diode can I replace this with?
Do I want to go with a Ge diode to keep the available voltage at a
maximum, or doesn't it matter?

Some random, small Schottky (for the low Vf) diode would probably do the
trick. I've seen germanium diodes used in similar circuits, but I
wouldn't recommend one with a lithium cell, as it will leak a small
current into the cell, which they don't like.

I've always thought that Schottky diodes were fine for the OP's stated
purpose, but after you raised the point about reverse leakage
currents, ISTR that Schottky diodes aren't too good in this regard.

- Franc Zabkar

Interesting.. I hadn't considered the reverse current aspect. I was
only worried about the voltage drop. Since you got me on the right
track I found this:

http://www.panasonic.com/industrial/battery/oem/images/pdf/Panasonic_Lithium_DesignMemoryBU.pdf

It gives a rough idea of the maximum allowable reverse current
depending on the application.

The examples in that document cite a reverse current of 4nA for a
Schottky diode and 6nA for a silicon diode.

But ...

A 1A 1N4001 diode has a typical reverse current of 50nA at 25degC:
http://www.onsemi.com/pub_link/Collateral/1N4001-D.PDF

A 1A 1N5817 Schottky diode has a reverse current of 1mA at 25degC:
http://www.onsemi.com/pub_link/Collateral/1N5817-D.PDF

The above leakages are measured at the rated reverse voltage, so at 2V
(5V - 3V) they would probably be a lot less. Nevertheless the figures
still suggest that Schottky diodes have worse leakage characteristics
than regular silicon diodes.

- Franc Zabkar

 

[Bob Larter]

http://www.panasonic.com/industrial/battery/oem/images/pdf/Panasonic_Lithium_DesignMemoryBU.pdf [...]

It gives a rough idea of the maximum allowable reverse current
depending on the application.

The examples in that document cite a reverse current of 4nA for a
Schottky diode and 6nA for a silicon diode.

But ...

A 1A 1N4001 diode has a typical reverse current of 50nA at 25degC:
http://www.onsemi.com/pub_link/Collateral/1N4001-D.PDF

A 1A 1N5817 Schottky diode has a reverse current of 1mA at 25degC:
http://www.onsemi.com/pub_link/Collateral/1N5817-D.PDF

The above leakages are measured at the rated reverse voltage, so at 2V
(5V - 3V) they would probably be a lot less. Nevertheless the figures
still suggest that Schottky diodes have worse leakage characteristics
than regular silicon diodes.

Interesting. I hadn't realised that 1N5817s were as leaky as that. The
OP should probably try to find a low reverse-leakage Schottky, or
substitute a standard diode & accept that the battery life will be
shortened due to the lowered memory voltage.

 

[Fox]

http://www.panasonic.com/industrial/battery/oem/images/pdf/Panasonic_....

[...]

It gives a rough idea of the maximum allowable reverse current
depending on the application.

The examples in that document cite a reverse current of 4nA for a
Schottky diode and 6nA for a silicon diode.

But ...

A 1A 1N4001 diode has a typical reverse current of 50nA at 25degC:
http://www.onsemi.com/pub_link/Collateral/1N4001-D.PDF

A 1A 1N5817 Schottky diode has a reverse current of 1mA at 25degC:
http://www.onsemi.com/pub_link/Collateral/1N5817-D.PDF

The above leakages are measured at the rated reverse voltage, so at 2V
(5V - 3V) they would probably be a lot less. Nevertheless the figures
still suggest that Schottky diodes have worse leakage characteristics
than regular silicon diodes.

Interesting. I hadn't realised that 1N5817s were as leaky as that. The
OP should probably try to find a low reverse-leakage Schottky, or
substitute a standard diode & accept that the battery life will be
shortened due to the lowered memory voltage.

--
    W
  . | ,. w ,   "Some people are alive only because
   \|/  \|/     it is illegal to kill them."    Perna condita delenda est
---^----^---------------------------------------------------------------

I found a compromise. A .7V 0.5 uA IR silicon diode with a 3.6V
lithium battery. I had a bunch of both lying around in my parts box.

 

[Bob Larter]

I found a compromise. A .7V 0.5 uA IR silicon diode with a 3.6V
lithium battery. I had a bunch of both lying around in my parts box.

That should work fine. ;^)