18 replies to this topic

[charlesc]

There are some extracts of a post pasted below, which came from a 'Get The Best reception' thread.
Discussed were some important aspects of DTV reception, along with some clear misunderstandings, and points that could confuse a reader.  It was buried away under a general post, so I brought it out here where it may be more useful.

The post is here, under "Get The Best Reception Regional Nsw, Which transmitter and which antenna".

alanh, on Feb 2 2005, 10:17 PM, said:

Please note this is general information. Use an antenna installer who can measure and error rates of digital signals (They should be able to give you Bit Error Rates (BER) figures of less than 8e-4.) If they cannot do this, get someone else.

"of less than 8e-4"
bellotv gave a pretty detailed set of examples of why the above comment could be confusing.
There is another aspect to this which bellotv also highlighted a good few years ago in another post.  This relates to using a figure that is the cut-off for what error figure an installer can give you.

"If they cannot do this, get someone else"
It isn't always a situation where the best can be given.
For some jobs (I'd imagine in particular remote country ones like belotv seems to have) you may not be able to achieve, as an installer, what you'd like in an ideal situation. But the customer may be happy with what you can give them.
Or the cost to deliver better and better signal error margin becomes increasingly high, and commercial reality and job cost take over as the main issue.

alanh, on Apr 28 2010, 01:52 AM, said:

BelloTV,
Values of errors I have quoted are prior to any error correction ie Channel BER or aBER or Pre-Viterbi. The value quoted is from Australian Standard 1367.
MER is the error ratio after all error correction and is not quoted as a ratio but in dB instead.

So BER will give the true level of errors. Post Viterbi BER is the errors left after the first attempt at error correction using the Viterbi Error Corrector. The signal is now Reed Solomon error corrected. The MER shows how well the error correction system can remove the errors.

So the antenna installation should be adjusted to minimise the channel BER. Once this has been achieved the MER should be >25 dB

MER and BER are both key measurements to be taken.  
In the DVB-T system, their measurement take-off point is essentially the same.

Modulation Error Ratio (MER) is not measured "after all error correction".
And it doesn't ... shows how well the error correction system can remove the errors.
This shows a fundemental misunderstanding of how the DVB-T system operates.  And what some of the key measurements are to try and make sure there is a reliable digital signal there.


Discussions of MER in this area of Australian TV reception (what the forum is about) would lead to QAM constellation diagrams and symbol representation.  And there is plenty of good reading available on that.

In simple terms though, real world signal impairments in the transmission and reception of the broadcast signal mean that the symbol landing points (when thought of in terms of a constellation diagram) aren't as precise as we'd like.  A certain 'fuzziness' occurs.

MER is a measure of how much difference there is.
The vector difference between the ideal target symbol vector and the transmitted symbol vector.  Expressed in dB.

It is measured after the digital signal is recovered from the analogue transmission.  And before it is passed on to other components of the receiver that handle the Forward Error Correction (FEC), specifically the Viterbi and Reed Solomon error correctors.

EDIT: Change to title to better reflect post purpose

Edited by charlesc, 03 May 2010 - 09:07 PM.

[James T Kirk]

Hi Charlesc

I think I have offered my bit on this before somewhere in this forum but happy to do again.
MER is one of the tools that may be used in the field however I believe it is sometimes misunderstood as to what it is actually telling the installer.

MER Modulation Error Ratio is intended to tell the transmitter operator how well the transmitter is able to reproduce the DVB signal at the transmitting site and is normally in the order of 35dB or better.

Once the signal is radiated to the viewers, an MER measurement if taken at this point will indicate the same figure as seen at the transmitter if the level of noise is very low, if noise is present the MER that is displayed will be below this 35dB figure. (Note many installer meters are incapable of resolving 35dB MER)

The point I would like to make is that once this 35dB figure is no longer being achieved in the field what is actually being measured is the MER + Noise and the noise is becoming quite significant. What you find is when you get say a figure of 27dB MER for example, the carrier/noise ratio if measured would also be around 27dB, note also at this time at the transmitter it is still pumping out an MER of 35dB or better, nothing has changed except the receiver has noise superimposed on the signal. So installers are actually measuring a carrier/noise figure in dB on their MER metering.

So MER measurements in the field are still a useful tool for the installer but is not a measure of the quality of the signal being transmitted. It is reasonable to expect that if a low gain receive antenna yield an MER figure of 26dB then a higher gain antenna would provide a better number as a consequence of getting the signal further above the noise.

James

[charlesc]

James, I do recall you posting some time ago on this.  
Thanks for again highlighting the difference in MER measurements taken from your end of things (the transmitter side) as compared to what we see at the reception end.

There was a point relating to this that I considered putting in the first post above, but wanted to keep that one short(!).

Sometimes you see reference to 'measurements' on installation meters that have MER scales that might indicate over 30dB.  On the meters that I use (Fracarro T40 Plus and AP-01) the MER range is typically to 26dB.
And most of the 'good' readings on those meters will top out at 25dB.
Reference meters for transmission monitoring applications do indeed cover the higher values.  But we probably couldn't afford them (or need them)!

When most impairments start to show on those meters, the 25dB lowers to say 24 to 22.  By the time you are in the 21dB or less range, there are major signal problems.  And the BER (and the CSI for that matter) clearly indicate this.
So the working range typically encountered is between say 19dB and 25dB.

In the real world of digital reception, measuring above this accurately is either not particularly relevant, and/or becomes much more expensive for the metering to handle (with good accuracy).
And in any case as you note, the signal once it has left the transmitter will experience many other 'impairments' along the way.  So received MER will not approach transmitted MER.

High MER measurements are taken (and expected) in the transmission part of the process, not the reception one.

[charlesc]

James T Kirk and stump_1100,
I think my topic title has been a bit confusing, so I've edited it to better reflect what the post went up for originally.
I wasn't suggesting we as installers could measure MER and somehow use that value to make any meaningful appraisal of the transmitter side of things.

Statement(s) were made in another post (referenced above) that MER was measured in the receiver after error correction.  And that it was a measure of how well the error correction system could remove the errors.
That is not correct.  So I posted as above.


Your responses above have highlighted some very important points though, so thank you very much for contributing them.  There clearly is a big difference in the measurement and handling of MER at the transmission and receiving ends of the system.


When I make measurements of MER, I am assuming the transmission side of things is working 100%, or at least, very close to.  This may not always be the correct assumption, but hopefully it is for a very large part of the time.  You'd know better the actual figures.

So when I see poor MER figures on-site, I'm taking that to mean there is a problem that has been introduced over the transmission path and/or at the receiving end.  
And I try to improve what is there by good antenna selection and positioning, signal distribution system improvements etc.

And these readings of MER are taken in conjunction with the other relevant readings of Bit Error Ratio (BER), Digital Channel Power (DCP) and Channel State Information (CSI).  Noise Margin simplified readings are also useful.

[James T Kirk]

Hi Charlesc

Genuinely I did understand what you were after from the start judging by your later comments, however my explanation may have fallen short.

Speaking specifically on issues at the receive end, when an installer is measuring MER in combination with other available parameters and they are getting MER under 30dB they are in reality just getting a carrier/noise figure, it's not MER but it is useful. When you take this "MER" number and make improvements to it by antenna selection, positioning and controlled distribution losses, it is being achieved by either and increase in received signal level, lowering noise or both.

Your reference to another quoting "BER limits that if not achieved, get another installer" and questioning the accuracy of this is completely valid. It can in fact be all there is at a particular location and if such a marginal signal is handled properly by the antenna installer can provide stable pictures.

For me getting best possible BER and receive level/channel power being at least 15-20dB above receiver RF threshold are all important.

I agree MER certainly is not affected by Viterbi or Reed Soloman error correction.

James

[bellotv]

charlesc, on May 3 2010, 09:43 PM, said:

When I make measurements of MER, I am assuming the transmission side of things is working 100%, or at least, very close to.  This may not always be the correct assumption, but hopefully it is for a very large part of the time.  You'd know better the actual figures.

So when I see poor MER figures on-site, I'm taking that to mean there is a problem that has been introduced over the transmission path and/or at the receiving end.  
And I try to improve what is there by good antenna selection and positioning, signal distribution system improvements etc.

And these readings of MER are taken in conjunction with the other relevant readings of Bit Error Ratio (BER), Digital Channel Power (DCP) and Channel State Information (CSI).  Noise Margin simplified readings are also useful.

Yes ,totally agree with you charlesc .

I also use MER as part of a full set of measurements to arrive at an acceptable result  and like you hope that the "boys on the hill"are keeping their end OK.

MER to me is like James T Kirk  suggests,an indicator of the signal to in band noise which is important to me as I have a lot of installs that are in terminally weak areas and require masthead amps to get the signal level up before hitting the loses.
MER also raises the flag to things like RF modulators in  E.G VCRs  on /near digital channels where daisy chains are used in a domestic installations.

MER has on numerous occasions showed up faulty amplifiers.

MER fluctuations are helpful in evaluating sites for minimizing wind/tree movement issues .

Charlesc ,I am sorry I can't comment on how MER in an instrument is derived as per your OP but perhaps some-one else knows more about whats in them .
All I can say is that MER in a Unaohm T40a is definitely a measurement that is useful and relevant to those on the receiving end .

Edited by bellotv, 03 May 2010 - 11:07 PM.

[beeblebrox]

Charles,

For those that can't read between the lines.....

"DON"T BELIEVE EVERYTHING YOU READ THAT IS POSTED ON A FORUM"

The world is full of well meaning individuals that can sometimes help out but when it comes to a professional doing their job if they are properly trained and qualified and experienced they will always give you a better result than a web forum can.

PS very much agree with your use of MER and it is a factor in any installation as part of the wider measurements on any install I use too.

Edited by beeblebrox, 03 May 2010 - 11:17 PM.

[James T Kirk]

I'm sorry if I may have given the impression that I consider MER at the receiver end to not be of value to the antenna installer, that is not so, my intention was to state what I believe MER is actually an indicator of when measured at the receive end.
When it comes to making a difficult location work, I am a an advocate of using all info at your disposal.

James

[bellotv]

Today I had a classic example of how helpful MER was.

Panasonic TV had bad pic brake up on most channels and no signal on another .

Channel power at the outlets were all up around the 60-70 dBuV. ( there was also a four way splitter feeding various outlets) so it wasn't a level issue

Some of the Ch bers , although low, were acceptable to the point where I would have expected so see an error free pic.Some were bad.Meter loosing lock.

But the MER on all channels was way low. MER 12dB -MER 16dB  on some   and MER 19dB on others.

I had spied a masthead amp on the way in and this was my first guess based on the above readings .A quick look at the full spectrum backed up my theory .( I've had this several times before)Even on a low resolution T 40a  I could see the spectrum was full of crap generated by the masthead amp.

A new one fixed it .( original was MHU34g ,generally a very reliable amp)
But since I had traveled a long way to this job we ( me and the customer) agreed that  I should also replace the aging 20 year old antenna (although it was still functioning it was quite rusted) .I also moved the antenna to a slightly better position that had more even CH power and better CH ber .

Edited by bellotv, 06 May 2010 - 09:21 PM.

[dig2all]

tad confused by bellotv's post.  you start talking about mer and finish with ber!

[charlesc]

dig2all, on May 8 2010, 05:32 PM, said:

tad confused
...  you start talking about mer and finish with ber!

Seems OK from here.
The first part relates to why MER can be handy, showing up a problem with the amplifier.
Then he also replaced the antenna and relocated it.  This gave better signal strength and quality.

The post shows practical useage of a few of the key digital measurements in various scenarios.

[dig2all]

someone mentioned the relationship between carrier to noise and mer.  in the early days most installer meters measured carrier to noise (in various ways) and used a look-up table to estimate mer.  the c/n estimators can be found out by adding noise or an interfering carrier within the channel in db steps (maybe tricky without some special gear) and seeing that the meter follows.  true mer will follow, estimated will not.  mer estimators are a bad deal today.

[dig2all]

true mer is limited by the range of measurement of installer meters.  they measure mer above 25db with the old c/n look up table, in this range they are all over the shop.  that makes installer meters (not the meters broadcasters use that cost a truckload or 2) that measure beyond 25db mer frankly dishonest.

[dig2all]

installers typically use low cost meters which do well to measure to + or - 3dB and often can be worse.  typical received mer range is between 10db and 20db so the useful range for installers is only 10db.  if that also varies by + or - 3db, that is a substantial part of the useful mer range that must lack precision.  so comparing mer measures between meters (say on this forum) has a high chance of error.

[dig2all]

ber is an count of errors (1, 2, 3, 4, etc.) that does not suffer the same potential for distortion.  it is more complex with 3 related measures but they better reflect activity within each receiver, and to start they need pay attention only to their pre viterbi cber.

[dig2all]

the 'industry' trainers who push mer as the measure could be said to be taking the path of least resistance, raising a generation of installers who only know mer on 'their meter' when ber is a better measurement that most meters have anyway.

in europe, ber is known as the king of measures for digital tv.

[beeblebrox]

dig2all, on May 9 2010, 05:14 PM, said:

the 'industry' trainers who push mer as the measure could be said to be taking the path of least resistance, raising a generation of installers who only know mer on 'their meter' when ber is a better measurement that most meters have anyway.

in europe, ber is known as the king of measures for digital tv.

In reality BER is the king of measures here too.  Most experienced installers will use MER as a guide but will always be looking at DCP and BER both cBER and PVBER.  
I found that the MER I got on the Unaohm wasn't as reliable as BER but since I moved to the Ikusi/Rover it seems more accurate at showing impairments with MER that are reflected in BER (which happens to be on a different screen on my meter).

I would note there are quite a number of installers only looking at Pass/Marginal/Fail on their meters which is better than no meter at all (plenty of those still running around) but these are the ones that very often cause the most grief for customers because they are often not in the best position and are close to the cliff.

[bellotv]

dig2all, on May 8 2010, 05:32 PM, said:

tad confused by bellotv's post.  you start talking about mer and finish with ber!

Woops,should have finished up with mentioning that the final MER was 24.1dB-25dB on all channels.

The whole point of the post was to suggest that the MER readings when interpreted with all other readings can help an installer identify a cause of problem.(in this case a faulty masthead amp)
I ended up on the BER theme as an aside to the main point I was trying to make but as I said it was  CH BER that I optimized by repositioning the antenna

I  am hearing you  regarding others running courses "pushing MER importance" ( This point was also raised to me by my local REP that works for the same Mob you work for)  

Obviously CH BER is King.There is no doubt .
But MER is its seconder .I'm sure you have see the power point by  Mark B showing the importance of in band noise being measured by MER. (You know the one with the Fraccaro single channel amp etc .)

What I was showing was that my Unaohms T40a meter  has a practical application of that in the field .In my case the" in band noise" was caused by a faulty Masthead amp.
I can't comment on how other meters that measure MER perform under the same conditions as I haven't had others to compare against .