Rega IO Amplifier Teardown

[amirm]

I recently reviewed the Rega IO amplifier. It did not perform well and got very hot. Members asked for a teardown and with kind permission of the owner, here it is: (click on image for 2X size):

Two ultra long screws is all that holds the thing together going from the back to front. Keeps labor costs low in assembly. Pray for me that they have a channel they go into so putting it back together won't be a nightmare!

AC mains from IEC terminal on top right is routed to the front panel for the power switch and then back. I would have liked to see some routed areas along that path to keep chances of arcing to a minimum.

On top you see the power supply rectification and filtering. That is fed with a toroidal transformer rated at ± 22 volts at 2.7 amps (from memory). The filter caps are secondary brand "Samhwa." We could forgive that but not the rating of just 85 degrees C as opposed to 105C. Considering how hot that side of the case was in my testing, I don't know how long these caps will last. Then again, with such height restriction, they probably did not have the option of getting a higher rated cap which would likely be larger/taller.

The main question was around source of the heat. I measured the left and right sides as being the hottest in the review:

As far as I can tell, there is nothing mounted on them. The power stage is those 5-pin transistors mounted to the bottom of the case. Could it radiate this much heat to the sides? Or are the rectifier diodes adding to it?

I made a light attempt at taking the tiny heatsink off them but they are screwed on tight. I did not want to force anything given how close to the edge this unit is on the thermal side. The transistors are quite large though so likely rated for good power handling.

I was surprised to see a number of through-hole transistors in there. Wonder what was special about them that they went through the expense of using them instead of surface mount ones.

I really liked the captive ring spade connectors for speaker terminals. They were screwed on tight with what seemed to be molded channel to hold them in their orientation. This should keep them from coming lose.

Overall, hard to find faults with the design of the unit given the heavy constraints of a compact unit like this using inefficient class AB amplification topology.

------------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

 

[GimeDsp]

I request you fry some bacon and eggs on it.

 

[JSmith]

I'm disappointed to see the poor measurements and heat management of this unit... was considering buying one at one point, so quite glad I did not. I would have thought Rega would provide a better amp than this considering the build quality and success of their turntables. It seems that possibly Rega should stick to what they do best in that respect.



JSmith

 

[YSC]

I request you fry some bacon and eggs on it.

Sounds like a plan, it must be a musical breakfast then

 

[pma]

Looking at the component placement and wiring, there is no wonder that @amirm has measured so many mains related spectral lines. Thanks for the review and teardown.

 

[DanielT]

I'm disappointed to see the poor measurements and heat management of this unit... was considering buying one at one point, so quite glad I did not. I would have thought Rega would provide a better amp than this considering the build quality and success of their turntables. It seems that possibly Rega should stick to what they do best in that respect.



JSmith

Sutor, ne ultra crepidam. A related English proverb is "A cobbler should stick to his last".[3

Ne supra crepidam - Wikipedia

en.wikipedia.org

Edit:
Then this What Hifi reviewing. Damn you have to take blood pressure medication if you read WHAT Hifi. Also applies to things other than Rega IO that they can waffle about. If it's not cables, it is, ... God knows what. Sorry for the OT. Just needed to get it out of my system.

Regarding Rega IO, it should be Whats NOT Hifi, by the way.

 

[pma]

I have added some remarks on the layout.

 

[tomtoo]

I wonder how long this thing is thermal stabel at 4ohm max. power?
10 minutes? Does it have thermal protection?

 

[pma]

I wonder how long this thing is thermal stabel at 4ohm max. power?
10 minutes?

For a class AB circuit, 1/3 of the maximum power is usually the worst operating condition re thermal stress, because the power dissipation is divided between the external load and output transistors (transistor Pc = Vce*Ic). At maximum power, there is less Vce voltage drop across the output transistors so the thermal load is not that high. 10-15W continuous would be probably the worst scenario for this Rega IO amp.

 

[Mart68]

Marketing to Engineer - 'Here's the case we're going to use, your job is to cram a 30 watt amp into it. We don't care how you do it or what compromises you have to make.'

 

[DSJR]

I repeat, the Io does NOT run hot like this in normal use. I suspect prolonged partying would result in early onset distortion as this amp is far too meek for long term high level use. It's not intended for 'big' systems and as I said in the other thread, larger Rega dealers go for the Brio R as the starter model and only see the point of this one as part of the 'little' record playing system they've put together.

 

[pma]

Marketing to Engineer - 'Here's the case we're going to use, your job is to cram a 30 watt amp into it. We don't care how you do it or what compromises you have to make.'

Yes. And I am so grateful that for most of my career in the last 30 years I have been independent of corporate style decisions and orders.

 

[KSTR]

Let's hope it has a thermal sensor that shuts the thing off when the case reaches 60°C.
We can see that the chassis bottom is an extruded aluminum profile which is an efficient heat spreader. All of the bottom and the side walls of the case will have similar temperatures. But the total area for radiation is small and this form factor provides close to no convection cooling.

Overall we have to agree, the engineers did a relatively good job to squeeze everything into that cabinet... given that the original design idea was to use a PWM amplifier and SMPS but they've encountered problems and switched to linear supply and class-A/B output in a late decision, at least that's what I read from someone's post in the original thread.

Given that @amirm has the clearance to open the unit, I suggest moving the mains transformer outboard (leads are long enough when cable ties are removed) to see how much it changes hum/buzz spectrum. The speaker wiring forms a huge receiving loop antenna and that alone can be enough to contract hum/buzz. The crosstalk may also be induced mainly by the speaker wiring (as long as load current flows, that is) because the left and right hot wires are bundled together (coming from the speaker relay. And the path to there -- the third, red wire -- from the channel farther away also is an open door for interference). So, I'd expect placing the transformer at some distance showed differences for hum/buzz.

 

[DSJR]

This is what just over double the UK price buys you (Elex R which is now £949 in the UK) - much more in keeping visually to peeps here I suspect - it's heavy too with solid casework. Remember the dealer takes a working percentage of the selling price.

 

[Lambda]

'Here's the case we're going to use, your job is to cram a 30 watt amp into it. We don't care how you do it or what compromises you have to make.'

"But use this famous class AB Design form the 80s ... customers liked the sound of it... and absoultly no SMPS customers hate them and it needs to be heavy so they feel the quality"

 

[restorer-john]

I have added some remarks on the layout.

I could add some remarks too, but it's ultimately just a really nasty design from end to end. Rega should be ashamed. But then again, they've never really cared one iota for measured performance, real world usage, or aesthetics.

 

[audio2design]

Looking at the component placement and wiring, there is no wonder that @amirm has measured so many mains related spectral lines. Thanks for the review and teardown.

It's like they sent the schematic to their Asian MFG partner for layout and this is what they got!

Look where the speaker gounds are .. way over at the power supply capacitors which will have a much different power supply reference, judging by the layout, from the small signal ground (need to see back of PCB). Will be different for each channel 2 as confirmed in measurements. Obviously not layed out by a competent analog engineer.

Star Ground? We don't need no stinking star ground!!

 

[Vini darko]

Volume pots right next to mains transformer. Just why?

 

[audio2design]

I was surprised to see a number of through-hole transistors in there. Wonder what was special about them that they went through the expense of using them instead of surface mount ones.

Not unusual in China to have some parts only available in through hole still if traditionally used in through hole assemblies. They already had to do a wave solder and placement so additional cost would be minimal wrt selling price even with a large markup. Good chance those parts were placed automatically.

 

[audio2design]

Volume pots right next to mains transformer. Just why?

Some old designer stuck in the 80s who insisted on a manual motorized volume control as opposed to digital volume control that could be placed by channel where needed saving much needed space and helping the abyssmal layout. Forest for the trees thing. That or a dumb marketing decision.