DIY Audio Projects Forum

What would you think happen when you feed your compartor output v4 into the micro?
With regards to maximum voltage on the micro pins

Theres not much point in making a micro controler to do the job anyways.

Using a multimeter with medium to high accuracy then doing V=IR will get you the current.
or simply adjust untill the two valves have same votlage accross cathode resistor.

Essentially your measuring the votlage with micro (the same thing), but you also introduce extra tollerances due to op amp and the micro ADC conver.
Its going to be less accurate unless you have atleast 10 bits ADC conversion ideally 16bits.
And anything more than 8 bits is massive pain in the ass to convert and get you head arround it.

You have the 8 bits stored in one variabble and the 2 bits stored in another variable for 10 bits.

And once you got the voltage set it will drift overtime, so zero point in extreme accuracy.

Maybe - yes, maybe - no. I doubt the 25Ohm pot is good for signal, there is no protection if one of the KT88 is gone for some reason either.
Auto-bias can be done without any harm if it won't be connected all the time.While Ref is referenced to the ground the amount of output voltage will be the difference voltage from the current sensing resistor (1Ohm x 0.09A = 90mV). The mistake here: the INA828 does not have rail-to-rail output, so it cannot produce such low voltage at the output.Yes if the amplifier is used by an "electronic person" and the person has a lot of time to check the bias.Not quite shure of your point. Any multimeter has some precision. Let's have 8-bit conversion of, say, 1Ohm current sensing resistor voltage (90mV) with "digital bias" of 200 of 255. 90mV/200=0.45mV or 0.45/2=0.225mA accuracy. You can dump the 25Ohm pot completely and put the correcting bias to the KT88's grids by a DAC. The "measuring circuit" can be connected to the current sensing resistors by a relay, so it starts to work only, say, in the moment of switching off. The sensing part could be made, say, with AD629, which do not need any 1M resistors to achieve high common mode voltage. And so on, so on...

Hi, Good thoughts however, some things that would seem to be potential causes of problems really have never materialized. The 25R pot is no worse than a resistor in the circuit and is (as long as you use a decent one) trouble free. It is a simple solution that does what is needed. This was a primary goal in the design of the amps. It is true that one output tube could fail and the other be exposed to damage. Fortunately KT88s and and any of the other various tubes that might be used are rather rugged and will tolerate the full current for a while. Several minutes at least and the degradation in the audio signal is a good clue something is wrong. Anyone who continues to use a amp that is clearly having problems really ought to buy something cheap that is solid state and pitch it when it goes tilt IMO. BTW a very large percentage of push-pull designs are equally subject to this sort of problem. Any that use cathode biasing with both tubes sharing the same resistor or other device will have the same problem. Considering the large number of Oddwatt amps in various sizes that are out in the world (both diy and kit) the number of issues is quite small. Well below 1% for all causes. (soldering problems are the number one by a large percentage of them) I'm pretty sure about the reliability as I personally answer all inquiries (for diy and kits). Since I consider my time valuable I don't want to spend much of it on problems.

As an aside.... there is a way to get away from the cathode 25R control. It does however slightly complicate the amp. Essentially you take a small amount of positive voltage from one of the clean locations (like the driver anode circuit) and feed it through a divider/splitter network to a pot (CT is grounded) of around 50K and split it off on each end to high value resistors (values vary by tube type) to each grid. If you set the voltage swing between the two tubes to be about 5-7 volts (grid to grid) you can adjust the balance that way. The drawback is it increases the voltage on the CCS and there are limits to what they can handle. You could use a negative voltage of similar size but it will tend to reduce the voltage across the CCS and possibly on peak signals push it into the non-linear region and cause distortion. Plus you need to develop a clean negative voltage first. I do the positive voltage version in some Poddwatts to allow for a slightly higher voltage on the CCS which gets it further away from the non-linear region and allows for about 20% more un-distorted output. It matters with the EL84 amps but is not significant in the bigger ones. The increased voltage across the CCS in the Poddwatt versions is well within the CCS ratings.

Good listening
Bruce

_________________
Some of my DIY Tube Amplifier Projects:

• PoddWatt: 5751 / EL84 Push-Pull DIY Tube Amplifier Project
• OddBlocks: KT88 Push-Pull DIY Tube Amplifiers (Mono Blocks)
• ForeWatt: ECC802S (12AU7 / ECC82) DIY Tube Preamplifier Project
• GrooveWatt: Tube (Valve) RIAA Phono Preamplifier Project

You still need a electornic dude to use the eqipment, needing to know excatly where to hook up the meter. (Hint all amps are diffrent)
So no, your not making life much easier for unexprienced people. And is likeley harder than just multimeter.

The refrence shifts the output by a dc bias(FIG 49), it dosen't limit the maximum voltage swing.
Sure you have some feedback from pins1 8 to set the gain, but there is no limit on maximum otuput voltage and is unacceptable design with regards to reliablity.
If any transient voltages on op amp turn on the micro will cook, even if the chip is designed to rise from 0V you cannot turst that it will do that every time.
The micro will burn feeding more than 5v into the pins.

Also good luck, you don't have any externally connected pull up or down resistors.
You can't rely the fact there may be internall pull up or low resistors.

Also you are not increacing the accuracy by using op amps, rather sense the voltage directly with the micro and then manipulate the data.
Also you have not shown the connection scheme for ADC voltage refrence, So I assume your using the micro 5v vcc voltqage as the refrence which is standard, So minimum ADC resolutiuon is 5/256 = 0.0195V which is quite large (19.5mV)
19.5mV of voltage error is unacceptable for a system like this.

You've also indirectly implied the voltage your sensing as is 90mV, so you will have massive error sensing something that is 0 to 300mV inputted to the micro.

Also good luck with keeping out noise using the 1Meg and 1Meg resistor divider to input the voltage into micro.

You made a wrong assumption that I'm the initial author of the schematic. My idea was different (and followed the schematic): no external equipment to connect to the amplifier (reasonable making life easier for unexperienced people), plus a way to discover any potentially devastating problem and to take a protective measures to prevent the damages.
Yes, but there cannot be more than 180mA in the circuit (limited by design with the help of CCS). So the maximum voltage would be 180mV. OK, if you insist - it is the addition of a penny diodes to prevent the overvoltage happens.
Let's not assume anything. As I pointed earlier - the schematic does have some mistakes, but it cannot prove the idea is bad.

No problem at all while the opamps are really buffers (amplification = 1) or have a small amplification. There is also an integration circuit, damping all AC noise from some frequencies upward.

There is zero load resistance at V4.
Hence V=IR
R is unkown or verylarge, more precise is the micro pin input resitance which is very large.

Hence you can get 24V at the micro input regardkess of the 10k serires resistance used.

No, the input is connected to the low impedance output of an opamp. There is no way the 24V will be here.

Then theres no point in connecting anything to it. If you made the micro pin an output.

Your not inputting any informatiion to the micro by having the V4.
Your also not outputting anything.

If its low umpedance output, so OP AMP will allow large currents to flow, so it will be able to deliver 24v.

The working principle of the INA828 is duplicate the exact difference from the two its inputs (V+, V-) to the output in relation to the reference wire. So "allow" and "will' in the case means different things.
I think if there is an interest for the module we can move our discussion to the dedicated thread and come to the precise solution.