Author Topic: Why HP power supplies doesnt have a obvious feedback loop?  (Read 2085 times)

0 Members and 1 Guest are viewing this topic.

Offline Efe_114Topic starter

  • Contributor
  • Posts: 46
  • Country: tr
Why HP power supplies doesnt have a obvious feedback loop?
« on: March 13, 2019, 12:33:15 pm »
I cant see the feedback loop even in a single hp power supply anyone can tell me? I am trying to build a power supply like the E3610 and i also looked at the others but cant find the feedback loop. Schematics are attached. (Also this thread will be a nice source for HP power supply schematics.)

Edit: are the attachments empty or is it just me?

Another edit: fixed by removing some of the attachments.
« Last Edit: March 13, 2019, 12:57:56 pm by Efe_114 »
Test gear: Aneng8009, 30V 5A chinese PSU( 1.5V peak-peak noise)
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #1 on: March 13, 2019, 01:49:35 pm »
You can't see feedback?
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: Efe_114

Offline tombi

  • Regular Contributor
  • *
  • Posts: 163
  • Country: au
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #2 on: March 13, 2019, 01:56:59 pm »
First the voltage control loop -

So in the first schematic (E3610A)  U1 is the voltage error amplifier. The non-inverting input is connected to either the output or the sense terminal (there is a switch) and the inverting input is connected to a pot fed from the +5 supply. The pot allows a variable voltage to be selected at the inverting input. The opamp will drive the loop to make its inputs equal.

U1 controls the main pass transistors (Q1, Q3, Q6, Q7) via a driver Q2.

So the loop goes Output/Sense -> U1 -> Q2 (driver) -> Pass transistors -> Output.

Current is measured by U4B - the outputs of U1 and U4B are connected via diodes to the driver transistor (Q2) so that whichever one is at a lower voltage controls Q2. This way the voltage control loop is driving the output until the current limit is exceeded and then the voltage control loop governs the output.

Does this help?

Tom
 
The following users thanked this post: Efe_114

Offline Efe_114Topic starter

  • Contributor
  • Posts: 46
  • Country: tr
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #3 on: March 13, 2019, 03:18:55 pm »
You can't see feedback?
Thank you! I tought S+ Was referanced to ground because the ground was marked S+ somewhere but i dont remember now. But how there is a feedback through the diode(CR8?) there?
First the voltage control loop -

So in the first schematic (E3610A)  U1 is the voltage error amplifier. The non-inverting input is connected to either the output or the sense terminal (there is a switch) and the inverting input is connected to a pot fed from the +5 supply. The pot allows a variable voltage to be selected at the inverting input. The opamp will drive the loop to make its inputs equal.

U1 controls the main pass transistors (Q1, Q3, Q6, Q7) via a driver Q2.

So the loop goes Output/Sense -> U1 -> Q2 (driver) -> Pass transistors -> Output.

Current is measured by U4B - the outputs of U1 and U4B are connected via diodes to the driver transistor (Q2) so that whichever one is at a lower voltage controls Q2. This way the voltage control loop is driving the output until the current limit is exceeded and then the voltage control loop governs the output.

Does this help?

Tom

Thanks! This really helped me but why are the back to back diodes(CR6, CR7) there between non inverting and inverting inputs? Wont they limit the voltage between 2 inputs to 0.6V?

If these questions are just so stupid you know who to blame for not making more fundementals friday.

EDIT: As i said S+ is referanced to ground in the display and referance supply
« Last Edit: March 13, 2019, 03:23:57 pm by Efe_114 »
Test gear: Aneng8009, 30V 5A chinese PSU( 1.5V peak-peak noise)
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #4 on: March 13, 2019, 03:42:46 pm »
Thank you! I tought S+ Was referanced to ground because the ground was marked S+ somewhere but i dont remember now. But how there is a feedback through the diode(CR8?) there?
If higher voltage is back-fed into the unit then that diode will conduct and the circuit will raise the voltage to meet the outside voltage and prevent damage to the circuit.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: Efe_114

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #5 on: March 13, 2019, 03:51:30 pm »
An opamp operating in its linear region has effectively no voltage between the inverting and non inverting inputs (that is to say is has a VERY large open loop gain and that this condition is maintained by feedback).

In normal operation the current sense amplifier drives its feedback loop to maintain almost no voltage between its inputs, however in the event of an overload or startup excursion the two diodes clamp the voltage difference to avoid some of the weirdness that can happen with opamps if you force the inputs too far apart (In this opamps and comparators differ, comparators are designed to run with significant differential mode input, opamps not so much (There are other differences).

Regards, Dan.
 
The following users thanked this post: Efe_114

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #6 on: March 13, 2019, 04:20:24 pm »
The circuit is indeed not that simple. Considering the "S" signal als part of the feedback loop is confusing. CR8 is only there for transients so that part should not be part of the loop.  One also has to take into account that the +5 V are relative to "S".

The better way to look at the circuit is to consider "S" as ground or the reference point. So the regulator is more like a negative side regulator.
The feedback loop would than be through R37, R13 , U1, CR5 ( that is conducting when in CV mode) Q2 and the power FETs  (acting as an inverting amplifier) and than through the raw power source (e.g. filter cap C2 for AC).
 
The following users thanked this post: Efe_114

Offline Efe_114Topic starter

  • Contributor
  • Posts: 46
  • Country: tr
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #7 on: March 13, 2019, 06:05:40 pm »
An opamp operating in its linear region has effectively no voltage between the inverting and non inverting inputs (that is to say is has a VERY large open loop gain and that this condition is maintained by feedback).

In normal operation the current sense amplifier drives its feedback loop to maintain almost no voltage between its inputs, however in the event of an overload or startup excursion the two diodes clamp the voltage difference to avoid some of the weirdness that can happen with opamps if you force the inputs too far apart (In this opamps and comparators differ, comparators are designed to run with significant differential mode input, opamps not so much (There are other differences).

Regards, Dan.
I just realised i know the opamp rules but i just didnt realise this fact. Yes i know the diffrences of opamps and comparators. opamps are more like analog devices made for generally linear operation and blah blah blah (am i wrong?). So diodes are there to prevent the diffrences between two terminals doesnt rise much so it will act as a comparator trying to work as an opamp and will oscillate badly right?
Thank you! I tought S+ Was referanced to ground because the ground was marked S+ somewhere but i dont remember now. But how there is a feedback through the diode(CR8?) there?
If higher voltage is back-fed into the unit then that diode will conduct and the circuit will raise the voltage to meet the outside voltage and prevent damage to the circuit.
Thanks!
The circuit is indeed not that simple. Considering the "S" signal als part of the feedback loop is confusing. CR8 is only there for transients so that part should not be part of the loop.  One also has to take into account that the +5 V are relative to "S".

The better way to look at the circuit is to consider "S" as ground or the reference point. So the regulator is more like a negative side regulator.
The feedback loop would than be through R37, R13 , U1, CR5 ( that is conducting when in CV mode) Q2 and the power FETs  (acting as an inverting amplifier) and than through the raw power source (e.g. filter cap C2 for AC).
Actually most of the older HP/Harrison power supplies are negative side linear regulators with very similar sink style Opamp(or amps built out of plain transistors if the design is very old)controllers driving a driver Pnp transistor and that transistor connecting the base of the Npn series pass transistors to -12V when needed. Also that Npn series pass is also connected to 12V via a resistor(generally 1.2Kohm)normally.
I couldnt upload manuals of most of them because they are over 1000Kb


THANKS TO ALL!
Test gear: Aneng8009, 30V 5A chinese PSU( 1.5V peak-peak noise)
 

Offline jaycee

  • Regular Contributor
  • *
  • Posts: 206
  • Country: gb
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #8 on: March 13, 2019, 06:28:59 pm »
The trick is that the opamp's supply rails are biased around S, which is connected to the main regulator's positive output. This trick is used in a lot of supplies, including the cheap Mastechs.
 
The following users thanked this post: Efe_114

Offline Efe_114Topic starter

  • Contributor
  • Posts: 46
  • Country: tr
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #9 on: March 13, 2019, 07:42:28 pm »
The trick is that the opamp's supply rails are biased around S, which is connected to the main regulator's positive output. This trick is used in a lot of supplies, including the cheap Mastechs.
What is advantages/disadvantages of this method? Why is this used in this design? I am a knowladge addict sorry for so much questions. Lemme guess, is it so OpAmps will never get out of supply voltage?
Test gear: Aneng8009, 30V 5A chinese PSU( 1.5V peak-peak noise)
 

Offline duak

  • Super Contributor
  • ***
  • Posts: 1048
  • Country: ca
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #10 on: March 14, 2019, 04:39:44 am »
Most of the hp lab power supplies were designed using the same basic idea of a floating regulator.  The output voltage is determined by a command voltage that is derived from a reference current passing through a variable resistance or just from a voltage that is provided from some external source.  This allowed them to be to be programmed externally, paralleled to increase current, put in series to increase voltage, and to sense the load voltage to compensate for voltage loss in the cabling.  This makes them very versatile and comes in handy when the supplies are used in automatic testing.  I didn't appreciate how brilliant the idea was until I started using the supplies.  Some of the Lambda and Kepco supplies use a similar idea.

Here's a link to a document that describes the idea: https://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1962-07.pdf

Here's a later version that goes into more detail: http://hpmemoryproject.org/an/pdf/an_06-65.pdf

Cheers,

 
The following users thanked this post: Efe_114

Offline jaycee

  • Regular Contributor
  • *
  • Posts: 206
  • Country: gb
Re: Why HP power supplies doesnt have a obvious feedback loop?
« Reply #11 on: March 14, 2019, 11:05:07 am »
What is advantages/disadvantages of this method? Why is this used in this design? I am a knowladge addict sorry for so much questions. Lemme guess, is it so OpAmps will never get out of supply voltage?

Well it simplifies design greatly as the opamps do not have to see the full supply voltage, which means a wide range of opamps can be used. It also means the supply voltages for the opamps can be well regulated and low noise, increasing PSRR of the main regulator.

It also scales very well.. for example it's no problem for the main output to be 18V, 30V, 60V, whatever you need... without having to modify the regulation loop much.
 
The following users thanked this post: Efe_114


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf