Bipolar PI design

[Detectorist#1]

For controlling of CD4053, maybe the simplest way, as said Carl, transistor translator have to be used

[ivconic]

• The micro must be powered from VCC and GND, not GND and VEE. _______________________________________________cor rected in post #25
• The sampling switches should be 74HCT4316, not 4066 (HC might work, I don't have any to test).____________________________corrected in post #25
• If you use a 4053 for IC1, use the 74HCT variety.__________________________________________ ___________________ no comment
• I need to add an input limiter to the 555 VCO.______________________________________________ ________________waiting to be done

... And, yes, Q1,Q2 can be any PMOS, I used the 9630 because it is common and cheap.________________________________Detectorist#1's post #27 is probably the better choice

Edit: the 74HCT4053 has a max supply of 10V so it won't work at 12V. And the CD4053 (max 20V) needs a 70% VIH drive, or 8.4V @ 12V. So this will likely need an additional translator.

I just checked, the 74HCT4316 is easy to find and the price is good.

[ivconic]

Hey Carl, no one holds anything against you, don't get me wrong!
Writing a book of almost 700 pages is no joke, it is a very tiring and difficult job.
Even more, I know from previous (own) experiences; it's great that mistakes happen and mistakes exist!
Because this is the ideal way to completely "enter" the project and to understand the whole story better!
Mistakes "wake up" a person and motivate him to think more deeply.
A project without mistakes is not something that benefits learning.
I often make 40 mistakes in a text of 20 simple sentences!
In a book of 700 pages, a dozen benign mistakes is an unprecedented success!

[Carl-NC]

As I said before, I need a few days to thoroughly go through this. Apparently what happened is that I built a first PCB proto of the circuit (derived from the PI-GB circuit) and found some problems with it. I made corrections to the circuit but then forgot to transcribe those changes to the book schematic, and also never ordered a final PCB to double-check the design. It was an oversight, which I blame on the fact that the PI section was some of the earliest material I "completed" and just forgot about the state of this circuit.

Anyway, here is a partial list of the issues:

• The micro must be powered from VCC and GND, not GND and VEE.
• The sampling switches should be 74HCT4316, not 4066 (HC might work, I don't have any to test).
• If you use a 4053 for IC1, use the 74HCT variety.
• I need to add an input limiter to the 555 VCO.

I say partial because I'm not done yet. And, yes, Q1,Q2 can be any PMOS, I used the 9630 because it is common and cheap.


Edit: the 74HCT4053 has a max supply of 10V so it won't work at 12V. And the CD4053 (max 20V) needs a 70% VIH drive, or 8.4V @ 12V. So this will likely need an additional translator.

[ivconic]

The least "painful" solution to this is to replace the charge-pump from the schematic with some DC-DC.
Of course; 7660 immediately comes to mind, "free-running" without external synchronization.
Apart from the voltage on the 7660 being limited to 10v... there is something else that makes me not like that solution, I can't see clearly what at the moment... but I feel it can be done better than that.


This is no longer valid.

[ivconic]

Hi ivconic,
Sorry to interfere with your project, but just look at this:

Feel free to "interfere" as much as you want.
The idea is to join as many people as possible here and to make something out of this.
Welcome!
It's not my project, but Carl's from the book.

I know you're right, but that's secondary now.
Read the post above, we have a much bigger problem right now!
Your post helped me discover what I wrote, thanks!

This is no longer valid.

[boilcoil]

About 4066 I'm not sure...
Thanks!

Hi ivconic,
Sorry to interfere with your project, but just look at this:

[ivconic]

Come again! Oooops!
Wait!
I think we have something here!
Carl pay attention to this!
Boilcoil provoked me to observe this more carefully!
If the micro is powered from ground to negative rail... and the negative rail is "generated" via charge pump controlled by micro....!???
How about that!?!

This is no longer valid.

[ivconic]

According to the diagram, only signals in the Vee - mass range will pass through 4066, i.e. you will lose the mass - Vcc part. In order to use the full signal, you, maybe, have to remove the direct control of the 4066 from the microcontroller, and add elementary drivers for control signals from the PIC, to cover the range Vee - Vss .

The micro is powered from ground to the negative rail, which is perfectly fine.
About 4066 I'm not sure... Carl must be involved, after all he's the author, not me.
But today a friend who does not speak English gave me similar complaints via email.
This is no longer valid.
...
Thanks!

[ivconic]

Hi Ivconic,

Max Uds on Q1 and Q2 is below 30V (12V supply + negative Ugs of IRF740). No need to use IRF9640 with max Uds=200V as upper transistors. 30V P-ch transistors haves many times lower RdsON and lower input, output and reverse transfer capacitances. Also the price is also reduced. AP9435GH (TO-252 case) and IRLML5203PbF (SOT-23 case) are suitable parts. IRF740A is a new version of IRF740, with more modern technology, with two times reduced output capacitance and many times reduced reverse capacitance (good for fast PI MD). IC7C have to be market as part of TL074 - not LM324, but this most probably is typing error.

I'm not sure if I read the same thing in Carl's book or on some forum, these days I browse a lot of similar things on the net and it's all mixed up in my head. But I think you're right.
P-channels can be replaced without problems.
As for the new IRF740A... that's very good information, I didn't know that!
As for IC7C; I just corrected that in the same post.
Because it is unnecessary to repeat the same things in several posts.
I know from previous experiences that some members get involved later in the topic and are then confused by the frequent repetition of schematics...
Thanks!

[boilcoil]

According to the diagram, only signals in the Vee - mass range will pass through 4066, i.e. you will lose the mass - Vcc part. In order to use the full signal, you, maybe, have to remove the direct control of the 4066 from the microcontroller, and add elementary drivers for control signals from the PIC, to cover the range Vee - Vss .

[Detectorist#1]

Hi Ivconic,

Max Uds on Q1 and Q2 is below 30V (12V supply + negative Ugs of IRF740). No need to use IRF9640 with max Uds=200V as upper transistors. 30V P-ch transistors haves many times lower RdsON and lower input, output and reverse transfer capacitances. Also the price is also reduced. AP9435GH (TO-252 case) and IRLML5203PbF (SOT-23 case) are suitable parts. IRF740A is a new version of IRF740, with more modern technology, with two times reduced output capacitance and many times reduced reverse capacitance (good for fast PI MD). IC7C have to be market as part of TL074 - not LM324, but this most probably is typing error.

[ivconic]

Any comments? Any mistakes?

[ivconic]

To make things clearer here is the revised schematic (click to enlarge):

[ivconic]

I understand what the confusion is. And now (I hope) the Source and Drain of all transistors are facing correctly: