UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
 
Thread Tools
Old 25th Apr 2018, 4:43 pm   #21
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

[QUOTE=astral highway;1037848]

Quote:
Originally Posted by Argus25 View Post
But it's great for plotting the Hysteresis or B-H curves of magnetic materials, All that you need is a function generator or AC source (sometimes an amplifier) and a current sense resistor on the primary, the voltage across the resistor drives the X amplifier. Then a simple OP amp integrator on the secondary to drive the scope's Y amplifier.
Here are some preliminary displays without the OP amp integration. The 'scope is integrating in this case, although I don't have enough data to interpret the results.

If I did, I could calculate H.

I used a ferrite toroid (homebrew gate transformer under test) such that N1=N2=16turns. It's the same giant toroid I've been discussing earlier and in other recent posts. Cross section is 200 square mm.

There is a simple precision 1R current sense resistor in the primary (X amp) and the Y amp reads the secondary.

-------
I don't know - and I'd need to - the source impedance of the signal generator, which is complimentary emitter follower which I built with a self-biasing arrangement.

See photo (4). It's a Texas Instruments idea, pretty neat. I get pulses of around the same current as I was previously, when I was running a conventional class B amplifier using the same pair, only then I had a few watts of heat to deal with. The Texas Instruments circuit appears to be close to 100% efficient - not even a smidgeon of heat. I'm not using BAS16 diodes, but Schottky diodes instead. (I ran this from 120KHz to 1MHz with the same little driver transformer and the results were phenomenally tidy and quite fast <70nS) square waves at the output into the load. The RMS drive current was only 30mA.)

-------

As you can see, the X and Y amplifiers are set to different sensitivities. When I set them to be equal, I displayed a vertical line.

I know that if I had the various missing unknowns, I could work out H from:

H=Vr*N/R*lc

where lc is the magnetic circuit length, and the result is in A/m.

The frequency of most interest to me is between 180KHz and 230KHz; however I ran the thing down to 65KHz or so as the hysteresis pattern changed in a non-linear way to show this distinctive balloon shape spinning on two poles, a little like our planet!

All of these (today's) tests are very rough-and-ready and actually have only a weak relevance to the development of my project. This is because the drivers I will be using in the end are bespoke high and low side gate drivers that can supply pulses over 12A each. They're on order.
Attached Thumbnails
Click image for larger version

Name:	hyst1.jpg
Views:	45
Size:	50.3 KB
ID:	161793   Click image for larger version

Name:	hyst2.jpg
Views:	32
Size:	48.5 KB
ID:	161794   Click image for larger version

Name:	hyst3.jpg
Views:	38
Size:	54.9 KB
ID:	161795   Click image for larger version

Name:	Self-biasingdriverIGBTbipolartransistor4-23 at 17.11.57.jpg
Views:	35
Size:	39.2 KB
ID:	161796  
__________________
Al

Last edited by Al (astral highway); 25th Apr 2018 at 4:50 pm.
Al (astral highway) is offline  
Old 25th Apr 2018, 11:26 pm   #22
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Al,

Those are interesting plots.

When I have done the B-H plots in the past I have used a sine wave generator to provide the magnetisation current, though in theory any shape would work because it is applied to the X deflection, but with high frequency square waves the rise/fall time could have Fourier components of well more than 20 to 50 times the fundamental. Then the test transformer's leakage reactance or inductance could have very significant effects reducing the current on the rise & fall times.

For most ferrite toroids you could start with a 10kHz sine wave and work up from there to a few hundred kHz and adjust the drive current upwards to see the effects.

With low currents, below saturation, the curve normally looks very much like an ellipse tilted over and with more drive toward saturation the tops and bottom tends to flatten giving a tilted over rectangular shape. Most of the cores I have tested have been iron, so I just did the test on them at 50Hz.

You could use that push pull follower to boost the output of a function generator it obviously would need a suitable power supply on the collectors and the transistors on heat sinks for sine wave rather than for switching use when the transistor dissipation would be much lower.

Last edited by Argus25; 25th Apr 2018 at 11:32 pm.
Argus25 is offline  
Old 26th Apr 2018, 5:13 am   #23
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Al,

On the device I'm typing on I can't reply in the usual way with quotes from a post, but with regards to your question on post 20, I think in this case all that happens if the OP amp's integrator's capacitor say is too large and the output voltage gets over integrated (at some test frequency) the amplitude of the integrated Y signal just gets lower in amplitude without causing any distortion.So excessive integration would just lower the amplitude (I think). The amount of integration required for a reasonable output signal would depend on the test frequency being used. An RC network (without need for an OP amp) also works ok for the integrator too, and has minimal distortion if the time constant is large enough.

Hugo
Argus25 is offline  
Old 26th Apr 2018, 1:48 pm   #24
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

I made a simple BH loop plotter! It works at 50Hz though in principle could work at higher frequencies too.

I used a small Variac (mains input) driving a step-down transformer to provide excitation current, and a couple of sense resistors (1 ohm and 100 ohm, switch selectable). Voltage across the resistor is proportional to H. Current flows through a winding of a few turns around the core under test.

For B, I used a separate winding on the core under test. This feeds an op-amp integrator (100k input resistor, 1nF feedback capacitor) which gives Vout = 10,000 x time integral of Vin - or putting it another way, 10V / milliweber. The long-term integral is zero of course, as it comes from a coil, but I still have a bit of trouble with the output gently floating in one direction or another due to offsets and 1/f noise, so I do have a rudimentary DC cancellation mechanism in place with a recovery time-constant of several seconds. No doubt one of Linear Technologies's chopper-stabilised op-amps would be the best choice here. (And yes, Hugo is right that a larger capacitor would just push amplitude down while leaving - with an ideal op-amp - the waveform unchanged in shape).

The H output and B output are fed to a 'scope in XY mode. I will hopefully put up a few photos later. Circuit is not available unfortunately but if pushed I will sketch out!

With a 1cm x 1cm core and 1 turn on the B-sense wire, the B output is 10V per tesla (as 1 tesla in 1cm² is 1mWb and we already know our integrator gives 10V/mWb). This is very convenient and gives me a good feel for saturation flux density in different materials, limited mainly by the accuracy I can measure the cross-section.

It is instructive to heat, or put pressure, on the core while viewing the loop. Ferrites can start to drop in Bsat not far above 100°C, while iron and nickel-iron are more immune (at least up to heat-gun temperatures).

Changing frequency would only involve squirting something else into the excitation winding, nothing else. I just haven't tried it. More voltage might be necessary, depending on the core's inductance, but that is a detail only. At higher frequencies I would expect eddy currents to start to kick in and broaden the B-H loop, with iron cores; with ferrite I would expect there to be little effect as losses are almost all hysteresis losses: the loop will stay the same shape, the spot which traces it out just whizzes around faster so more energy is expended each second purely by virtue of this.

Last edited by kalee20; 26th Apr 2018 at 1:49 pm. Reason: spelling
kalee20 is offline  
Old 26th Apr 2018, 1:57 pm   #25
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Quote:
Originally Posted by Argus25 View Post

You could use that push pull follower to boost the output of a function generator it obviously would need a suitable power supply on the collectors and the transistors on heat sinks for sine wave rather than for switching use when the transistor dissipation would be much lower.
Hey Hugo, I did put a suitable regulated voltage supply on the collectors. Curiously, the current shown on my DC bench supply is so immeasurably small that it registers as zero. This is despite the waveforms being way more defined! - see below. It's an efficient follower stage, that's for sure! In fact, it only draws current when I head below the frequencies where there is any meaningful field strength or flux density to display.


Quote:
Originally Posted by Argus25 View Post

For most ferrite toroids you could start with a 10kHz sine wave and work up from there to a f[/CENTER]ew hundred kHz and adjust the drive current upwards to see the effects.
I was suspicious that this particular toroid had a poor medium frequency response, so I cranked it (down!) to beyond 1MHz, and yes, my suspicions are right. We can clearly see that at 4MHz, the field strength displayed on the X axis is high, and the flux density on the Y axis is also much more defined. This remains true as we head down into the 3MHz, 2MHz (not shown) and 1MHz, although less so.

However, both field strength and flux density are much less defined at the frequency of interest to me, around 180-230KHz, and unacceptably so.

I'd never considered applying this test to check the suitability of this toroid as a gate transformer, but alongside the traditional test of square waveform purity, undershoot and overshoot, it turns out to be excellent.

Time to rewind on another core.
Attached Thumbnails
Click image for larger version

Name:	4MHz hysteresis.jpg
Views:	25
Size:	61.1 KB
ID:	161854   Click image for larger version

Name:	3.77MHz hysteresis.jpg
Views:	24
Size:	41.8 KB
ID:	161855   Click image for larger version

Name:	200Khz hysteresis.jpg
Views:	30
Size:	65.1 KB
ID:	161856  
__________________
Al
Al (astral highway) is offline  
Old 26th Apr 2018, 6:40 pm   #26
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

I have photos of the B-H gizmo, a couple of cores, and plots obtained. I took these this afternoon. These were made at 50Hz.

Cores are similar dimensionally and I used the same number of measurement turns. Scope setting was the same for each. Noteworthy is:

(1) Ferrite has much lower Bsat than transformer iron

(2) Area of loop is much smaller so losses are very much lower

(3) Although difficult to estimate, the near-vertical portions of the loop are similar in slope, indicating that inductance factor Al is similar

(4) the transformer iron, when it does approach saturation, has a much sharper 'knee' than ferrite (toroids actually have an inherent softening effect of the 'knee' because the inner circumference saturates first, saturation gradually spreading outwards - as the inner portions have a shorter path length)

If instead I ran at 100kHz or more, I'd expect the ferrite to look fairly similar, and the steel to look like nothing on earth. I have looked at B-H loops at SMPS frequencies in the very distant past, and found ferrite tends to open-up as frequency increases (eddy currents start to kick in, because although resistivity is high, the material also has significant permittivity so it looks like a distributed lossy capacitance).

The job that Astral wants his CT to do, as long as B is kept well below 0.3T (and this is simple to calculate), the CT can be considered relatively blameless.
Attached Thumbnails
Click image for larger version

Name:	FullSizeRender.jpg
Views:	58
Size:	84.8 KB
ID:	161866   Click image for larger version

Name:	FullSizeRender (1).jpg
Views:	47
Size:	62.5 KB
ID:	161867   Click image for larger version

Name:	FullSizeRender (2).jpg
Views:	47
Size:	64.0 KB
ID:	161868   Click image for larger version

Name:	FullSizeRender (3).jpg
Views:	44
Size:	101.6 KB
ID:	161869  

Last edited by kalee20; 26th Apr 2018 at 6:47 pm. Reason: additional info
kalee20 is offline  
Old 26th Apr 2018, 7:45 pm   #27
Ed_Dinning
Dekatron
 
Join Date: Jan 2004
Location: Newcastle upon Tyne, Tyne & Wear, UK.
Posts: 8,172
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Hi Peter, looks like a useful tool for sorting out core types.


Al - Have you seen the Mar/ April edition of Elektor. It has details of a scope current transformer using a LEM unit with quite a bit of circuit detail.
You may also like to look at Rogowski (think that is spelt correctly) current sensors.

Ed
Ed_Dinning is online now  
Old 26th Apr 2018, 10:12 pm   #28
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Peter: Thank you; neat work, too.

The 50Hz results are very interesting. You kind of suggest running at 100KHz; I’d be interested. More so, even at 180-250 KHz.


Would that be ok?


Ed: Thank you; I’ll definitely look up Elektor.

Also, Rogowski coils: I’ll take a look!
__________________
Al
Al (astral highway) is offline  
Old 26th Apr 2018, 10:58 pm   #29
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Kalee20,

Nice B-H recordings.

The ferrite shows minimal to zero coercivity and has negligible remanent magnetism (unlike the iron) and it follows practically an identical magnetisation-demagnetisation path. So that would make for a very efficient transformer as area inside the hysteresis loop represents energy losses per cycle.

I have always been very impressed by Electromagnetic Engineers of yesteryear prior to the days of silicon/IC technology where a lot of complex functions get performed with little effort. They really knew their core materials. Instead they had to get the best out of their magnetic materials. (I can't post the circuit as I'm away at the moment) I will post an ingenious circuit later, where the designers chose to make the cores for their SCR pulse transformers out of Alnico. Once a single pulse passes, the remanent magnetism disables the the transformer for one polarity of pulse. It will only work again when an opposite polarity pulse or reset pulse is energised on another winding to reset it. This allows a ring counter, like a shift register linked in a loop, to be created, all using the magnetic memory of the cores in the pulse transformers. This pre dates IC based digital logic, but it didn't stop the imagination of the designers from getting the same result that they would have with a chain of flip flops.

Also in dynamos it is important that the core material for the field windings has good remanent magnetism or they won't self excite. The core properties also add to the filtering effect so that the field coil switching voltages cause little ripple in the armature output. But remanent field can be easily reversed, when converting from + to neg ground.

Hugo.

Last edited by Argus25; 26th Apr 2018 at 11:08 pm. Reason: Typo
Argus25 is offline  
Old 27th Apr 2018, 7:19 am   #30
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

They were at their most impressive with magnetic amplifiers.

I once had to do some re-design of multi output SMPS where transductors were used to regulate the minor outputs. It was a fascinating field (sorry) to dig into.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 27th Apr 2018, 7:36 am   #31
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

.....I could have added, Kalee20, that is a real darling of a B-H tester you made.

I have a couple of small variacs from a project I didn't use. Using the example of your fine work, I think I might make a formal B-H tester too. I'm looking forward to seeing your circuit.
Argus25 is offline  
Old 9th May 2018, 6:29 pm   #32
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Quote:
Originally Posted by Argus25 View Post
I'm looking forward to seeing your circuit.
Here it is! Sorry for the delay, I've been away playing with horses.

The LT1055 is a precision FET input op-amp, it works well, but as the circuit is an integrator a microvolts or so of offset gives a scope display which slowly wanders up or down the screen. And, the direction and speed of movement can abruptly change - this is 1/f noise in the op-amp manifesting itself as a change in offset. I tend to think that a chopper-stabilised op-amp would be a better choice here, but as you see I added a very-low-frequency pole-zero pair to roll off the gain before it headed to infinity at DC.
Attached Thumbnails
Click image for larger version

Name:	image.jpg
Views:	111
Size:	60.3 KB
ID:	162519  
kalee20 is offline  
Old 9th May 2018, 7:21 pm   #33
Al (astral highway)
Dekatron
 
Al (astral highway)'s Avatar
 
Join Date: Nov 2006
Location: London, UK.
Posts: 3,496
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Hello Peter,

Thanks for posting this, it's a neat circuit and you've thought of everything.

I'm busy for the next couple of weeks but I do intend to run some high frequency tests for comparison when I get round to it.

Thanks for your contribution!
__________________
Al
Al (astral highway) is offline  
Old 10th May 2018, 12:08 am   #34
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Pulse tests of current transformer - clearly shows instantaneous rate current cha

Quote:
Originally Posted by kalee20 View Post
I added a very-low-frequency pole-zero pair to roll off the gain before it headed to infinity at DC.
Great little circuit and its easy to see how filtering most of the AC component off the feedback loop helps to maintain the frequency response of the OP amp over a wide range but also give it DC stability.
Argus25 is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 10:49 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.