A rotary encoder controlled integrated chip amp LM3875

[justblair]

Originally posted by banana
Is it the MAX5440 will be at -12dB position upon every powerup? That seems too loud.

Anyway, I like the idea of graycode encoder control without MCU. If there's a way to access the resistor ladder directly without using the build-in opamp, I'll like to try it too.

Interersting project.
Please show us the LED display looks like

Hi Banana

Yes the Datasheet states -12db on powerup, which does seem very loud. However I imagine that this will not be too much of a problem. There is a different feature of the chip called shutdown. When its pin is pulled low it goes into this mode. When returned from shutdown it remembers the previous pot position. Frankly I coudn't see how this differs from Mute. The power consumption of the chip is the same in either. So I haven't factored it in my design.

Unfortunately the resistor ladder can only be accessed via the op amp AFAIK. MAxim do other chips such as the DS1802 that dont include op amps, however the rotary encoder and LED feedback seems to only come on chips with op amps. I'm going to feed it a reasonable quality signal and hope the opamps are not too bad. The output looks fairly linear on the datasheet, but time will tell.

The LED display is simply 5 LED's. Its up to the user how this is implemented. Watch this space, I have a couple of ideas on this.

[justblair]

I have had an idea on the LED front.

Using PNP transistors is a good idea. However it would mean redesigning my board.

Not a problem, there is plenty of space in that area that could house the required transistor circuit.. When I was thinking about the redesign though, I had a brainwave.



If I remove the led power rails, resitors and replace them with a small header, I can take this part of the ciruit off-board..



This allows me to fold in the caps at the top of the picture, basically it shortens my board by around 2cm.

The other advantage of doing this for me is it allows me to then produce more varients at the LED stage. For instance I can have a very low parts count circuit for red/yellow LED's, or a more complex board for using higher voltage drop LEDs. I can also attach the encoder to the design as it is board mount.

On a standard photo etc prototyping board say 100mm by 160mm I can fit both the AMP pcb and several variations of the LED/encoder mounting board. this will allow me experimentation for less money.

On the subject of transistors. I am a bit stumped as to how to imlement this. Here is where I am so far...



Now on the resitors after the collector, I am looking at values of around 100R. However i am baffled as to how to calculate the resitors that limit the current to the Base. I have read up on tutorials for doing this, But each one I have found assumes the supply voltage ground and the base to be one in the same. In my case the base will connect to 0v via its resitor (well 0.4 according to the maxim datasheet) Could someone help me out on the maths here?

Other than that I did not get any work done till now, My GF whisked me away to Perth to see Jools Holland play on friday, and we hit the Edinburgh Jazz festival today. I dont mind that kind of interuption at all. Tomorrow though I need to get this design finished ready for etching.

Any help as always apreciated..

Blair

[BWRX]

Hi Blair. Your circuit schematic for the PNP switched LEDs looks good. Calculating the resistor values is really quite simple. You want the PNP transistors to act as switches so you want the base current to saturate the transistor. This will cause the Vce voltage to drop to around 0.2V. Using the Vce drop, the LED voltage drop, and the 5.5V across the rails, you can calculate the value of the LED current limiting resistor. Say you want around 10mA for each LED and the LEDs have a drop of 3V at 10mA. The value of the current limiting resistor should be about (5.5-0.2-3)/(0.01)=230ohms.

Since there will be about 10mA of collector current you just want the base current to be high enough to saturate the transistor. The minum DC current gain of a 2N3906 at Ic=10mA is around 100, so you want the base current, Ib, to be greater than about 10mA/100=0.1mA. Just to be sure the transistor is saturated you may want to shoot for Ib closer to 0.5mA. The open drain output of the chip will pull the base resistor to ground, so we can calculate the value of the base resistor to be about (2.75-0.7)/(0.0005)=4100ohms.

Kapeesh?

[banana]

Hi Blair,

How about using current mirror (current control current source), instead of common-emitter config to driver LEDs?

This should revive better the variable brightness, I guess

[justblair]

I think I get it, but let me feed it back in my own words...

The Vce drop is caused because the transistor when open is acting like a saturated diode, hence the 0.2v drop? The datasheet I am reading (On Semiconductor) gives



So lets make a guess then that at 20ma the voltage drop will be similar enough to 2.5Vdc

Now my blue LEDs have a voltage drop of 3.2 and 20-30ma so the calculation will be

(5.5-0.25-3.2)/(0.03) TO (5.5-0.25-3.2)/(0.02)=68R TO 102.5R

So 100R is correct for the resistor on the collector.

For the base, It is 20ma/100=0.2ma multiply by 5 and use

(2.75-0.7)/(0.002)=1025 or a 1K should be about right?

the 0.7 in the above calculation is the voltage drop from base to emiter?

Brian you are very patient by the way... Thanks for your help.



Does this mean that 5.5v is still not enough for the blue leds? Or can I run without the resistor

[justblair]

Originally posted by banana
Hi Blair,

How about using current mirror (current control current source), instead of common-emitter config to driver LEDs?

This should revive better the variable brightness, I guess

Hi Banana. i think that going with common emiter is better for me, as the part count is lower. And I dont want to go any further in the math than I have reached so far...

The variable brightness you mention?

Another issue that crops up there is the LED indicator from the MAX works in increments. The lowest LED's will light up to full brightness, the final one dims bit by bit before going out. This allows for more increments to be represented. I think I could bias the transistors to give this effect, but I will have to look into it.

I said htis earlier, though when I typed it I was tired. Looking at the datasheets more closely, this would appear to be true for the push button chip(MAX5486), but not the MAX5440. However the circuit will work with either the MAx 5440 or the 5486 if a trace is cut. The brightness is controlled by PWM.

I can make a sperate daughter board to control the LED's for the 5486. I dont know how the 35ns switching time will be affected by the transitors. I can give it a try, though the rotarty encoded version is my number 1 priority. I have samples of both chips however.

[justblair]

Been busy again today.

Made the changes to the main PCB, and a front panel designed incorporating the Rotary encoder and the LED's





The PCB I have managed to shorten considerably. This will allow me more space longitude wise in the case, It gives me more options in what is a tight space.

I did this by removing the LED supporting components off of the Main PCB. This also allows me more flexabilty in that I can add several differnent configs to one PCB and split it into the main PCB plus other front panels.

I will produce the first board with two options, one for a MAX 5440, the other for the MAX5486. So I can produce either a push button or rotary encoded design.

I also can factor in several options for LED placement with each board.

The first I have produced uses a 35mm aluminum knob on the encoder. The LED's are mounted behind the knob itself. So when lit up they will reflect light from underneath the knob. I should get a halo effect around the knob.

For the push button design, I expect to produce a straight forward line of LED's with push buttons at either end. Another option I may design uses a touch sensing chip instead of buttons.

The encoder I got cheap off of E-bay. I unfortunately dont have details on its use, so the multimeter will be out.

The encoder has 5 pins on it plus two ground lugs.



I am guessing that the pins denoted with A-C-B are the encoder inputs, the other two non labelled pins I am guessing will be the terminals for the push button selector.



I should find out tomorrow when I get a chance to test it with a meter. If not I will have to redesign the panel.

Anyone have any experiance with these?

[BWRX]

Originally posted by justblair
The Vce drop is caused because the transistor when open is acting like a saturated diode, hence the 0.2v drop?

A PNP transistor is called as such because its internal structure is back to back PN junctions (a single PN junction is a diode). Without going into any real detail, the transistor will try to minimize the Vce voltage with increasing base current. If the emitter is connected to 5V and the collector is connected to the positive lead of an LED whose negative lead is connected to ground, the collector will sit at the same voltage that is across the LED regardless of the current going through the transistor. If a resistor is in series with the LED the transistor will try to minimize Vce if you pull enough currrent from the base to saturate the transistor.

Originally posted by justblair
Now my blue LEDs have a voltage drop of 3.2 and 20-30ma so the calculation will be
(5.5-0.25-3.2)/(0.03) TO (5.5-0.25-3.2)/(0.02)=68R TO 102.5R
So 100R is correct for the resistor on the collector.

I bet those LEDs will be very bright at 20mA, so I would use at least 100ohms as you correctly calculated.

Originally posted by justblair
For the base, It is 20ma/100=0.2ma multiply by 5 and use
(2.75-0.7)/(0.002)=1025 or a 1K should be about right?
the 0.7 in the above calculation is the voltage drop from base to emiter?
Does this mean that 5.5v is still not enough for the blue leds? Or can I run without the resistor

0.2mA*5 is 1mA or 0.001A not the 0.002 you used in the above caculation.
You can short the current limiting resistor for maximum current if you find that the LEDs are too dim with the resistor in place. Most LEDs are very bright at 20mA though, unless you have special super bright high current, high intensity ones.

Originally posted by justblair
The variable brightness you mention?
I said htis earlier, though when I typed it I was tired. Looking at the datasheets more closely, this would appear to be true for the push button chip(MAX5486), but not the MAX5440.

I never actually looked at the MAX5486 datasheet, but the MAX5440 definitely does not have variable brightness. Either the LED is on or it's off.

[justblair]

Originally posted by BWRX

I bet those LEDs will be very bright at 20mA, so I would use at least 100ohms as you correctly calculated.

They are bright, however as they will be behind the volume knob, it is the reflected light that will be visible. i can experiment a bit with this to get just the correct levels.

Originally posted by BWRX

0.2mA*5 is 1mA or 0.001A not the 0.002 you used in the above caculation.

Doh...

Originally posted by BWRX

You can short the current limiting resistor for maximum current if you find that the LEDs are too dim with the resistor in place. Most LEDs are very bright at 20mA though, unless you have special super bright high current, high intensity ones.

I doubt that it will be necessay, but at least its an option. The effect i want will hopefully be subtle. Most listening will be done in the evenings.

Originally posted by BWRX

I never actually looked at the MAX5486 datasheet, but the MAX5440 definitely does not have variable brightness. Either the LED is on or it's off.

Yep I have to admit I blundered here origionally. I was working from both Datasheets. I want the design to offer either rotary or push button control. I assumed incorrectly that one chip was pretty similar to the other. I spotted my error today. La weekend i deinately overdid it.

However the design seems to be reaching the final stages. Just got to design the power supply PCB and have a general tidy up. That should be a lot simpler. I converted the schematic, and it pretty much came into a useable design when I ran the autoroute. Just have to compact it and get the placing right for the voltage outs. They will hopefully line up pretty closely with the boards inputs. That way I can minimise the cabling mess.

I didn't apreciate how much work would be involved in this design. However I am feeling pretty confident with it now. Lets hope I have not missed something major.

[justblair]

Getting there with the final design. Got the power supply laid out...





Going online tonight to place an order for parts. Give it a final once over and hopefully will be etching this week.

Please have a peruse and point out any blatant errors. I have designed the power supply pcb with all of the components mounted on the underside. The AC components are at the non signal end of my amp board, The Mur 760's will sit closer to the torroid as the power board is 120mm long as opposed to 85mm long. i have placed two of the Mur 860's at the very edge of the board. I will mount an alu plate to these which will serve as a ground shield to the amp board.

In mounting the boards "back to back" I will hopefully introduce as consistant a ground shield as I can between the power components and the amplifier components as possible.

It also means I can couple the heat producing components on the power to the base of the case, the heat producing components on the amp to the top. I have some heatsinks resued from some blown computer psu's which are the right design to do this for me.

Another advantage is that the wire connecting the boards together will line up producing the shortest run. Less messy cabling wise and good hopefully at avoiding adding junk to the DC voltages.



It's going to be tight, but hopefully not impossible. I can cut out unused bits of the pcb's if need be to allow rear connections etc

I have also lined up the regulators to allow me to use one continuous heat sink.