Technical questions to Harbeth's designer

[Ted Rook]

Hi Alan,

I'm curious to find out the type of relay you use in your development tool kit for switching between crossovers at speaker power level. The reason is that I created a little project recently, a pair of polarity inversion relays, and I am dismayed to find that the octal 10A relays I chose have a degrading effect on the stereo image. Now I am not a tweaker, I'm the last person to believe unscientific claims about connection wizardry, but, the thing I was trying to hear disappeared. I find absolute polarity to be close to the limit of detectability and then only with pure stereo mic recordings. So, I need a relay that doesn't impact the stereo image. What do you use? As you may recall I'm in the US but some relay types are generic. An RS or Farnell part number would be handy :-)

Best Regards

Ted Rook LS5/12 and now M30

PS congratulations for the resurrected forum. I wonder if I'm the only one who doesn't like being titled junior member!!!?

[A.S.]

Hi Alan,

I'm curious to find out the type of relay you use in your development tool kit for switching between crossovers at speaker power level. .... So, I need a relay that doesn't impact the stereo image. What do you use? As you may recall I'm in the US but some relay types are generic. An RS or Farnell part number would be handy :-)

I grabbed whatever was convenient. I don't recall if they were fancy, but I suppose given a choice I'd go for mercury contacts. Or gold. Or silver. Or just about anything with a low contact resistance ... down in the thousandths of an ohm range. Put that into perspective: a typical speaker cable could introduce 500-1000 times more resistance.

For a really good 'snap' action, use the full voltage specified for the relay (AC or DC, as specified) and select one with spare contacts that can be wired through an LED which give a positive, visual confirmation that the contacts have truly changed over. You don't want to go to all that bother and then kid yourself do you?

As for stereo image changes - I'm baffled. There *must* be some other reason: logic says that a relay can't impair image. It doesn't have a brain, so it doesn't know about such things.

[Ted Rook]

Thank you Alan.
Since posting my question I learned a little more about relay science and I found this paragraph particularly relevant......seems I was careless in my selection of silver contact relays for this application

quote from potter Brumfield application note http://relays.tycoelectronics.com/app_pdfs/13c3236.pdf

Also, an interface voltage of several tenths of a volt can result
with fine silver contacts because of the sulfide film. This film has been
known to capture and imbed airborne dirt. Breaking through this film
generates electrical noise. Because of this, fine silver contacts are not
used for low-level switching, such as audio circuits. Rather, fine silver
and silver alloy contacts are for use in circuits of 12 volts, 0.4 ampere, or
more.

"Back to the drawing board"

Ted

[A.S.]

Thank you Alan.
Since posting my question I learned a little more about relay science a....
"Back to the drawing board"

Ted

Er, I really don't think so Ted. It's all a matter of proportion. It's so easy in audio to lose that sense of proportionality.

I'd be willing to bet you that regardless of whatever caveats you read in that report concerning contact materials, that *all* that matters for the purpose of your speaker-switcher is a fairly low resistance - say a quarter of an ohm or so. Anything better than that isn't worth the bother. But I anticipate that you'll not believe me until you've spent serious time and effort: I find that time after time in this crazy industry.

Why do we chase after a degree of perfection that simply isn't achievable - or worthwhile - or even necessary? No wonder women are not attracted to this industry!!!

[Ted Rook]

Alan,

you are being very generous with your time thank you.

What you wrote is exactly the down to earth common sense reply that we need in audio. I bought a relay with "quarter ohm contact resistance" That demonstrates my devotion to common sense too.

There's only one problem, it obscures low level detail, you would hear it too if I could bring it to you and you would be as amazed as I am.

...........that *all* that matters for the purpose of your speaker-switcher is a fairly low resistance - say a quarter of an ohm or so. Anything better than that isn't worth the bother............

Relay makers go to great lengths to get the contact material appropriate to the varieties of end use: type of load, power rating, and I find, minimum values for current and voltage.

I now see I made an error in using the maximum power rating (10 Amps) to select the device for this application. I should have used the minimum signal as the defining characteristic, milliVolts and microAmps, and I am confident that there I will find the superior contact behavior required.

Ted

[A.S.]

Alan,

What you wrote is exactly the down to earth common sense reply that we need in audio. I bought a relay with "quarter ohm contact resistance" That demonstrates my devotion to common sense too.

There's only one problem, it obscures low level detail, you would hear it too if I could bring it to you and you would be as amazed as I am. ... Relay makers go to great lengths to get the contact material appropriate to the varieties of end use: type of load, power rating, and I find, minimum values for current and voltage ...
Ted

Maybe Ted, maybe.

I still don't believe that after the countless millions of solder joints, semiconductor joints, metal film resistors, capacitors and the like in the signal path from the mics to your speakers that right at the very final hurdle a bit of contact resistance (oxide or whatever) can make an audible difference!

[Ted Rook]

I too subscribe to the robust principal where the audio signal is concerned. It traverses millions of solder joints as you rightly say. If you did not have more important calls on your time I would send you the set-up for you to hear for yourself, maybe I should? You would need a quiet listening space, a CD made using pure coincident stereo mic technique and ten minutes to set up. I'm continuing with the project and I'll let you know what transpires.

Please do not think I have been taken over by some malevolent high-end audio energy, my feet are still firmly on (technical) ground :D

Ted

[A.S.]

I too subscribe to the robust principal where the audio signal is concerned. It traverses millions of solder joints as you rightly say. ...
Ted

Well you know Ted, I've just remembered how I coped with a conceptually similar situation when I built a change-over box to test in-line cables (CD to amp for example) about 20 years ago. In-line signals are, of course, much smaller that those fed to the speaker so, presumably, more vulnerable to contact issues?

What I did was to buy reasonable quality relays with many contacts that changed over. When energised, some opened, and some closed but with a momentary overlap, so there was no break at all in the signal. No gap. No silence.

Under those conditions, it was not possible to hear, see, smell, feel or in any other way deduce the change over. All cables tested passed this test as sonicly identical in every way. However, using conventional break-before-make relays, which introduced a small silent gap as they flicked over, there were "definite differences".

[John Parkyn]

Alan:

There is a fair amount of talk about "Radial".

What is "Radial" and what is so great about it?

Are there plans to use "Radial" in all Harbeth models ... I could be wrong but I didn't see mention of "Radial" when I scanned the Monitor 40 and HL P3ES-2 pages on the (main) Harbeth website.

Thank you,

John

[A.S.]

What is "Radial" and what is so great about it?... I could be wrong but I didn't see mention of "Radial" when I scanned the Monitor 40 and HL P3ES-2 pages on the (main) Harbeth website.
John

Research And Development Into Advanced Loudspeakers (RADIAL) is the Harbeth patented and exclusive cone material that resulted from a Harbeth inspired, British Government co-funded project. This Govt. incentive scheme was set up about 1990 as a helping hand to get engineering graduates into industry. The so called 'Teaching Company Scheme' was conceived to bring those latent skills to SME's (small-medium businesses) just like Harbeth.

It was underpinned by the ongoing project management (at the local University/Poly, head of dept. Dr./Professor level). The deal was that Harbeth would fund 50% of the three graduates salaries and sundry costs for the three years and 0% of the equipment budget. At the end of the project all the equipment would be inherited by the University. As some ?150,000 of taxpayers money was being spent, the Govt. inspectors attended the quarterly Review meetings at the University and grilled us all on progress through formal presentations before further funds were released.

The darkest moment was about 18 months in when we realised that the petrochemical companies were never going to support us with the supply of custom made plastic film (to enable us to vacuum form cones) because our lifetime usage was, in their scale of thinking, pitifully small. That forced us to think laterally - to vacuum forming - which, due to the vastly higher set-up (tooling) costs the Govt. people eventually agreed to co-fund. It was the spark that revolutionised the whole project.

So, RADIAL is an advance polymeric cone, using a blend of chemicals with just the right acoustic properties. I have started the write-up here but it is not complete: http://www.harbeth.co.uk/designersnotebook/chapter4/chapter4-1/index.php

Someone did ask once if this was an appropriate use of taxpayers money. First, the Govt. has received back in import savings, export earnings, payroll tax, corporation tax and VAT many times what they invested in us so for the taxpayer it was a spectacularly successful deal. E.K., our team leader is now a senior figure in the design of MRI scanners, so mankind daily benefits from his contribution. His success is offset by sadness: it's such a pity that Jane, one of the original team was killed in a car accident as I discovered when I needed to make a second large batch (500kg+) of RADIAL in 2004. Jane was the young process engineer who with her sleeves rolled up and wielding a huge spanner frequently dived into the extrusion moulder when the prototype RADIAL plastic mixes went 'off the boil'. Her tenacity and gritty, willing determination to tease into bonding materials that were on the theoretical edge of combinability (for reasons of temperature and density) was absolutely invaluable. We got there in the end.

One of the great benefits of a thoroughly peer-reviewed project is proper documentation - and E.K. and I were able to pick-up Jane's excellent and detailed notes (and especially those literally in the margins of her reports) and regenerate the ratios, pressures and temperatures exactly. RADIAL lives on.

The RADIAL cone is used in the M40's midrange driver. Picture here of the chopped RADIAL granules (which are naturally white) still warm out of the granuliser.

[mou]

I've found controversial information both from local dealers and the 'net. Some says it needs a long burn in period, some others says a few hours or even minutes, and, to make it all worse, some say that all units are burnt in prior to leaving the factory.
Where do you stand here and what's your opinion about the significance of run in time or burn in time on a loudspeaker unit?

[A.S.]

I've found controversial information both from local dealers and the 'net. Some says it needs a long burn in period, some others says a few hours or even minutes, and, to make it all worse, some say that all units are burnt in prior to leaving the factory.
Where do you stand here and what's your opinion about the significance of run in time or burn in time on a loudspeaker unit?

Um. You do know that this and many other questions are covered in the main website's FAQ don't you?! We update that regularly so please do check there from time to time.

http://www.harbeth.co.uk/faq/index.php#16

Burn-in is another one of those minor technical issues masquerading as a universe of misunderstanding and folklore. The only things that can age (hence, burn-in) are the suspension and the ferrofluid in the tweeter's magnetic circuit. The suspension reaches it's final value after a few hours 'working' (or even minutes if the signal is bass-heavy and/or quite loud: beware) and the ferrofluid become optimally runny after a few seconds or so.

After that, the only thing that changes is ..... YOU!

[A.S.]

Regarding cone movement and the harbeth grille frame .... there is a balance between how far the frame has to protrude from the front of the cabinet (which could look rather inelegant) and the danger of the cone (softly) brushing against the cloth. So, recently I made some experiments to see how much power I needed to push into my speaker to touch the cloth.

The mximum linear excusrion of a woofer is defined (approximately) by the thickness of the magnet top plate, which is what you would see under the corrugated 'spider' if it and the cone were transparent. It is solely resfponsible for folding the magnetic fiend into a narrow gap, and hence onto the voice coil. The voice coil is 12mm long, and the top plate is 6mm thick, so at rest, there is 3mm of coil sitting above the top plate and 3mm below. (This is all approximate due to fringing effects etc.).

So that means that is the bass unit (and of course it's voice coil) move more than 3mm towards you from the resp position or 3mm backwards into the cabinet from rest, the voice coil has partially left the linear magnetic field. So, plus and minus 3mm of linear movement is barely any motion at all, yet it will produce a good loud sound, so by definition, a fore/aft excursion that has the bass unit flailing about must be beyond ?3mm - hence with rising distortion, reduced power handling etc. etc.. Under such extreme excursions the contact with the cloth actually acts as a last-stop break.

Should you think that this is some peculiar limitation of Harbeth speakers I can assure you that the industry standard is 6mm thickness for top plates, so this would be an identical situation right across the (quality) speaker industry.

Summary: if the voice coil is not fully and continuously immersed in the magnetic field at all times, the magnet can not control the motion of the cone, and technical and subjective performance will degrade.

[danrubin]

That is the theory and specification. My experience was this: I was taking frequency response measurements of my system using the Rives Audio CD. At the low end (20Hz up to about 50Hz), I heard loud and distinct noise from the grille (buzzing, "rattle"). The woofer surround was making contact. This was at volume levels of about 75db at 1KHz, down 10-25 db at the lower frequencies.

So, I removed the grilles and listened to the Harbeths "naked" for the first time. It fixed the noise problem. Moreover, though perhaps a touch bright, I find I am preferring the sound. And thinking back on earlier exposure to Harbeth (CES 2004, where I heard them for the first time) and the great sound I heard from them at the Denver show last September, I believe the grilles were always off. Doesn't look as good, however.

In designing the SHL5, is the speaker "voiced" with the grilles on? If so, how "off" do you feel the sound becomes when the grilles are removed?

Another issue is finding a fix for the problem.

-Dan

[A.S.]

That is the theory and specification...

There are two factors here: first, are you sure at those low frequencies that it wasn't the port sucking the cloth as opposed to driver contacting the cloth?

I should have added that the frequncy response of the Harbeth SuperGrille is extremely flat - I'd hesitate to say the best in the speaker industry for a grille when most are just sheets of wood with an aperture (or two) chopped in them, covered with cloth and slapped onto the front of the speaker. However, part of the sonic excellence of the Harbeth solution is that the cloth is close to the drivers. As with all things there has to be some sort of balance between forces.

[danrubin]

are you sure at those low frequencies that it wasn't the port sucking the cloth as opposed to driver contacting the cloth?

I'm fairly sure, but I will re-run the exercise in the next few days and post again. My impression is that the driver was contacting the crossbar of the grill in addition to, or rather than, the cloth. I'll observe closely and report back.

[Gary DiChiara]

I have experienced a similar noise on my M40s (only when using a test CD and only at a certain frequency around 60hz) and upon closer examination I determined the noise was actually eminating from the grille cloth slapping the front of the crossbar. A few little pieces of blueTack between the grille cloth and the crossbar eliminated the noise completely. with musical programme, it was never a problem. Hope this helps.

[Ned Mast]

Apart from the wonderful manner in which my SHL5's reproduce recorded music, I am also quite impressed with the fine cabnetry (in my case, cherry-wood). Do you recommend that any type of wood preservative (some kind of oil?) be used (sparingly, of course) to maintain the fine finish of the cabinets?

With thanks,

Ned Mast

[kevint]

Hi Alan,
You mentioned the standard BBC stand is made from tubular Speedframe, would like to know more details about the usual BBC way on setting up monitor.
Do they use any spike for level adjusting or floor the stand?
Anything to interface the monitor to stand? Without any thing in between, it is often hard to get perfectly flat top stand to rest the monitor on.
Is the room carpeted?

You also mentioned Monitor series is designed for well-lagged environment, can you suggest ways to make domestic room more lagged?
Thanks!

Kevin

[A.S.]

Good question, but I think you would be surprised at the reality. In a professional sound organsation there is zero time and zero interest in tweaking and tuning. The speakers are tools to do a job; they are installed, and almost certainly never adjusted from the moment they are installed to the moment they end their life many long years later. They are not even given a dusting! So, forget all about spikes, cups, fancy cables etc. etc.! All that matters to the BBC engineers is that the sound from the speakers is sufficiently neutral and characterless that they don't have to get up from the mixing desk and make their way down into the studio (perhaps 100m away) to check that they can believe what the speakers are telling them.

Yes, the BBC control rooms (or cubicles as they are known) are very well lagged: thick 300mm absorbing panels on the walls; good quality carpets with good quality underlay (most important), double or triple glazed, and with absorbing ceiling tiles.

Of all these, I would say the easiest and best to aim for at home are ...

1. A really good quality (wool) carpet with the best quality fibre backed with rubber crumb underlay you can afford. Do not save money on the underlay: it is a false economy.

2. The thickest, softest lines curtains you can afford, floor to ceiling on the side walls. You can arrange these such that you can slide them open when you want to listen, and open them when you don't.

3. Book cases. As many as possible - and ideally (although very inconvenient) with the spines against the case and the open pages facing into the room.

4. If you have your own den where cosmetics are not important, then consider lining the walls with top quality fibre-on-rubber crumb underlay with the crumb against the wall. The SHL5 was designed in a small 4 x 5m room using just that on the walls and it works extremely well - although does make the room look dark.