[f] AMB on Facebook

ε24 Rebooted

epsilon24

Re: ε24 Rebooted

Postby sv511 » February 4th, 2016, 3:10 am

This is the complete power switch driver ready to test. I did not install a snubber diode.

I would prefer to have it mounted in a case with a fused power entry module before I put power to it. The case should be here this week, but I still have to do something about the panels.

I'm hoping to get access to a CNC milling machine at work so I don't have to wait for FPE
You do not have the required permissions to view the files attached to this post.
sv511
 
Posts: 71
Likes: 9 posts
Liked in: 10 posts
Joined: January 13th, 2016, 1:00 am
Location: Isle of Anglesey
Country: United Kingdom (uk)

Re: ε24 Rebooted

Postby passive420 » February 5th, 2016, 10:45 am

Nice little board. From here, it looks like it works ;)

Can I ask what relay you are using and how many amps you estimate you are switching? My build is a modest 250VA for the amps and 80va for DAC and output stage. I have a CRYDOM 2425-10 on hand.

Some reading on correct relays to use for high inductive loads would suggest non zero crossing relays (-10 suffix) and peak phase switching, an example of a phase switched relay below. Do you have any knowledge on these matters?

http://www.crydom.com/en/tech/newslette ... 0types.pdf

https://www.gavazzionline.com/pdf/RM1Cdatasheet.pdf
passive420
 
Posts: 23
Likes: 4 posts
Liked in: 0 post
Joined: January 17th, 2016, 2:45 am
Location: UK
Country: United Kingdom (uk)

Re: ε24 Rebooted

Postby sv511 » February 5th, 2016, 12:06 pm

The on-board SSR is LCA182S RS stock number 906-1133


The off-board SSR is a Crydom ED24-D5 RS stock number 740-1359

This is a zero-voltage turn-on relay. Maximum resistive load current is 5A and maximum peak current 600A for one 50Hz cycle.

I'm not an expert on SSRs. I'm hoping there won't be start up power factor issues. I'm building σ22 supply to be limited by with the 100VA transformer for peak loads and heat sink thermal limits for average load.

The plan is to use this to with a β22 driving headphones so everything is overkill.

5A * 240V = 1200VA. I will use a 2A slo-blo fuse which is more than enough for the 100VA + 1,5VA transformers. The relay should be fine but I won't really know until I try it. I should look at the data sheet for the fuse and make sure it will protect the relay. In rush currents large enough to damage the relay would probably trip the breaker on the ring main or blow the fuse in the plug.. (If any American's happen to be reading this, UK wall plugs have fueses in them) Again I'd have to check the data sheets.



I haven't modelled the start up in-rush current. 4400µF x 2 isn't that much although the ESR of a 2200µF 50V PW capacitor is a low 0.023Ω. I would guess the initial current will be mostly limted by transformer resistance. (Copper loss is just over 9% at 100VA) Spice model would be a good weekend project ...
sv511
 
Posts: 71
Likes: 9 posts
Liked in: 10 posts
Joined: January 13th, 2016, 1:00 am
Location: Isle of Anglesey
Country: United Kingdom (uk)

Re: ε24 Rebooted

Postby sv511 » February 9th, 2016, 7:40 am

I am using an Avel Lindberg Y236405 100VA transformer.

DC resistance in the primaries is 20Ω when wired for 230VAC. That alone should be enough to limit the in-rush current to safe levels for the SSR.


Transformers "reflect" the secondary impedance back to the primary. I can't predict what load the transformer will present to the relay but it won't be pure inductance or even primary inductance in series with 20Ω. With a σ22 on the secondary, it is far from obvious to me. The rectifier plus C5, C6, C7 and C8 are probably enough for an in-rush current model.

I may go ahead and plug this into LTSpice although the model will suffer because I don't know enough about the transformer. My Tenma LCR meter died so I don't have any way to measure inductance.


How to simulate a transformer in LTSpice:

http://cds.linear.com/docs/en/lt-journa ... lhardt.pdf
sv511
 
Posts: 71
Likes: 9 posts
Liked in: 10 posts
Joined: January 13th, 2016, 1:00 am
Location: Isle of Anglesey
Country: United Kingdom (uk)

Re: ε24 Rebooted

Postby sv511 » February 10th, 2016, 3:15 am

I did a first order simulation of σ22 in-rush in LTSpice as an exercise. I already knew it was not going to overload the relay but this puts some numbers behind it.

The model is the transformer secondaries, the rectifier and the filter capacitors. The rest of the regulator will do something but the overall effect on inrush should be small compared to the 4700µF capacitors. (see the σ22 schematic)


I am simulating the transformer secondaries as an ideal voltage source in series with an external 1 Ω resistor. I measured the DC resistances in the transformer

SchCap copy.png



The current in the primary will be about 1/4 of the current in the secondary. This is not even close to the relay's peak current rating.

The next step would be to model the transformer primary including a wild guess at the inductance.

Currents copy.png
You do not have the required permissions to view the files attached to this post.
sv511
 
Posts: 71
Likes: 9 posts
Liked in: 10 posts
Joined: January 13th, 2016, 1:00 am
Location: Isle of Anglesey
Country: United Kingdom (uk)

Re: ε24 Rebooted

Postby sv511 » February 10th, 2016, 7:22 am

I added a transformer model with my measured winding resistances and a few wild guesses for inductance.

With dual 2H primary windings, the inrush current peaked at 4.9A. This is only an approximation but the result is two orders of magnitude below the 600A peak limit of the relay.

Overload stress test: Putting 1Ω resistors on the output (yikes) causes a 3.5A RMS load at the primary. Shorting the outputs results in a 4.31A current in the primary. The relay will survive until the fuse blows.

WithTrafo copy.png
You do not have the required permissions to view the files attached to this post.
sv511
 
Posts: 71
Likes: 9 posts
Liked in: 10 posts
Joined: January 13th, 2016, 1:00 am
Location: Isle of Anglesey
Country: United Kingdom (uk)

Re: ε24 Rebooted

Postby anson-IT » March 8th, 2016, 7:04 pm

sv511 wrote:The boards came from PCBWay in Shenzhen China. I have been happy with the service from PCBWay. This is the third time I have used them and they have always shipped on-time. Their $13 for 10 boards up to 10x10cm offer is perfect for this sort of project. 1.6mm FR4, 1oz. copper and HASL may not be up to the audiophile standard but hey $13! 2oz copper, 2mm thick FR4 and other goodies are available but at a higher price.

I paid $25 for DHL express shipping from China even though it is nearly double the cost of the boards. HK post is much cheaper ($10) but takes about a month to get to Wales.


I used to use Olimex even though their design rules can be annoying. I could live with 8mil traces but some of the spacing rules were 14mil. The worst was Olimex charged for every drill change from their rack of 7 standard metric sizes. I used to spend days editing Eagle component libraries to avoid their drill rack chargers. €1.00 per drill change adds up.

Olimex stopped their PCB prototype service in September and couldn't find another inexpensive fab in the EU. The PCBWay boards are generally better than Olimex. Their design rules and capabilities are much easer to use. PCBWay $13 boards have a minimum 6 mil track and spacing and Olimex were 8 / 14. The best part is not having to worry about drill sizes. They also offer Green, Red, Yellow, Blue, White and Black solder mask at no extra charge. I tried red once. It made the traces hard to see. I think I'll try blue next time.

About the only thing better at Olimex was their boards all came with Immersion gold. That is not included in the $13 special but HASL is probably better for development. Gold is pretty to look at though :)


Reference: (if you care about such things)

http://www.pcbway.com/capabilities.html

https://www.olimex.com/PCB/TechnicalInfo/


could you tell me how is http://www.pcbway.com/b quality?i will choose one manufacturer for my project
front-end language is a little difficult
anson-IT
 
Posts: 1
Likes: 0 post
Liked in: 0 post
Joined: March 8th, 2016, 6:42 pm
Location: US
Country: United States (us)

Re: ε24 Rebooted

Postby dsolodov » March 9th, 2016, 8:55 pm

I did a variation as well using MAX6816 as documented in the thread http://www.amb.org/forum/power-controller-t2692.html?hilit=power%20controller#p24768.
Completed AMB projects: M³, σ11, σ25, ε24, γ1, γ2, LCDuino
User avatar
dsolodov
 
Posts: 362
Likes: 30 posts
Liked in: 4 posts
Joined: September 4th, 2010, 10:02 am
Location: Canton, MA
Country: United States (us)

Re: ε24 Rebooted

Postby sv511 » February 4th, 2020, 1:29 am

I found an error in Rev 0.4. The component library for the Dialight 553-0xxxF had the anodes and cathodes reversed. Rev. 0.5 includes the fix.

Image
sv511
 
Posts: 71
Likes: 9 posts
Liked in: 10 posts
Joined: January 13th, 2016, 1:00 am
Location: Isle of Anglesey
Country: United Kingdom (uk)

Previous

Return to ε24 power switch driver circuit

Who is online

Users browsing this forum: No registered users and 0 guests

Quick Links

AMB audio DIY resources
AMB audio shop

cron