Path: ...!weretis.net!feeder9.news.weretis.net!news.quux.org!eternal-september.org!feeder2.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: john larkin <JL@gct.com>
Newsgroups: sci.electronics.design
Subject: Re: switchmode gyrator
Date: Fri, 15 Nov 2024 07:33:18 -0800
Organization: A noiseless patient Spider
Lines: 94
Message-ID: <3opejj149flsl6jp0cdn65r49bi671844r@4ax.com>
References: <onquijt47loo1c3krbof1mpq0l16a8afqv@4ax.com> <rbj6jjt0tm8m6kjmnp7akojmshrhbo3uop@4ax.com> <evt6jjh8utojvc9ugmi54f0qp8aehfka93@4ax.com> <vh07fo$1nq2q$1@dont-email.me> <ktj7jjhg6is1vttu0vu8oqi966225b5qnm@4ax.com> <MnFZO.12180$xDBd.2237@fx13.ams4>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Injection-Date: Fri, 15 Nov 2024 16:33:21 +0100 (CET)
Injection-Info: dont-email.me; posting-host="d6d6c47e9e8f1fec4ab58e060dcc2cc8";
	logging-data="3658648"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+9DArgmKL37QfrzXFMXqsp"
User-Agent: ForteAgent/8.00.32.1272
Cancel-Lock: sha1:4/9/lBnfYEZs0C9WJAbIQzc1pcU=
Bytes: 4919

On Fri, 15 Nov 2024 21:28:58 +1100, Chris Jones
<lugnut808@spam.yahoo.com> wrote:

>On 13/11/2024 9:05 am, john larkin wrote:
>> On Tue, 12 Nov 2024 19:44:39 +0100, Jeroen Belleman
>> <jeroen@nospam.please> wrote:
>> 
>>> On 11/12/24 16:49, john larkin wrote:
>>>> On Tue, 12 Nov 2024 07:49:13 -0500, legg <legg@nospam.magma.ca> wrote:
>>>>
>>>>> On Sat, 09 Nov 2024 06:08:40 -0800, john larkin <JL@gct.com> wrote:
>>>>>
>>>>>>
>>>>>>
>>>>>> Inductors are awful. Their energy storage is worse than electrolytic
>>>>>> caps by about a factor of 1000.
>>>>>>
>>>>>> https://electronics.stackexchange.com/questions/326177/energy-density-comparison-between-inductors-and-capacitors
>>>>>>
>>>>>> One could in theory make a switchmode gyrator that would make a
>>>>>> capacitor look like a programmable-value inductor.
>>>>>>
>>>>>> I have an application for that, but it would take too much engineering
>>>>>> and runtime complexity to make it worth doing. I guess I'll just have
>>>>>> to buy a bunch of giant, heavy custom toroids.
>>>>>
>>>>> I'm sure that there are other reasons why an inductor
>>>>> will be used besides the lack of time/energy/resources
>>>>> to 'design them out'.
>>>>>
>>>>> Energy storage is just one means to an end.
>>>>>
>>>>> RL
>>>>
>>>> Sure. We want to design some dummy loads that will simulate relays,
>>>> solenoids, stepper motors, torque motors, with programmable R and L.
>>>> Seemed to me that using caps to make fake inductors would be a good
>>>> way to do that.
>>>>
>>>> It's at least an interesting idea to play with. Maybe we can
>>>> switchmode simulate R+L all at once. We would have to store energy and
>>>> dissipate power to do that.
>>>
>>> Relays and motors do not behave like simple inductors. For example,
>>> while a relay armature is moving, the back EMF is high enough to
>>> make the current _drop_ briefly. Modelling that requires more than
>>> a simple gyrator.
>>>
>>> Jeroen Belleman
>> 
>> My customer is building giant rackmount boxes full of heavy inductors
>> as part of his dummy loads. We want to replace them.
>> 
>> Given a generalized switching impedance simulator, I guess one could
>> model a DC motor.
>> 
>> I am considering a powered impedance simulator, not the theoretical
>> gyrator. Just sort of a gyrator.
>> 
>
>Many inductive loads can produce kilovolts when switched off suddenly. 
>If you build a switched-mode equivalent, unless it contains a large real 
>output inductor, your class-D output stage might need supply rails of 
>kilovolts to emulate the real inductor accurately. That sounds expensive.

An electronic load would of course have specified voltage, current,
and power limits.

But some of the simulated inductance - a few mH maybe - could be real
inductors, and they would fly back a bunch. We do need to accept a
customer PWM drive and sort of behave right.

I don't think that kilovolt flybacks are common in real areospace
systems. That could damage wire insulation and interfere with other
stuff. 

>
>If the load you are emulating contains a TVS or other clamp as part of 
>the load, then it would be easier to emulate, but you wouldn't be able 
>to emulate the fault condition where the clamp fails open.
>

Yes, the customer's drivers have flyback diodes or equivlent, or just
a continuous low-Z PWM drive. We can't simulate every corner case,
lightning strikes or major system faults.

It's just an interesting problem, simulating a pretty high power
inductor without a hundred pounds of iron. What I'd never thought much
about is how terrible inductors are at storing energy.
Inductor:capacitor:supercap:battery are roughly 1000:1 steps in energy
density.