For coating 1 to 10 micrograms of mAb can be used. Since you are testing the antibody I would recommend 1, 2, 5, and 10 micrograms. I suggest you to keep the antigen and the primary antibody. You can not have too many variables in a single run which needs a lot of samples to process. After deciding on the coating antibody, you can test antigen concentration and primary antibody concentration. Goodluck 

So capture antibody concentration range from 10-1000mg?

The capture antibody I use is purified 

In the link below it is recommended to use 100ng/ml? it must be primary Ab concentration.

http://www.abcam.com/protocols/antibody-dilutions-and-titer

Exactly too many variables. But I would like to know a typical range of concentrations for capture, sample and primary Ab used in sandwich ELISA although the same concentration would not be applicable to my experiment. If anybody has come across such paper I would like to read through it. I'm sure somebody must have done this before so it would be great if they share their insights.

I edited my answer. Please check it. if you are using purified antibody the range is 1 - 10 micrograms per ml. The below link provides the concentration of the antibody for coating.

http://www.abcam.com/protocols/sandwich-elisa-protocol-1.

And the link you provided suggests the primary antibody concentration for the purified antibody.

The sample concentration to be used is highly variable, which depends on the type of the sample. I would suggest dilution range from 1:10 to 1:10000 if you do not know the concentration of the antigen in your sample. If it is a purified protein, you can use the range between 10 ng/ml to 10 micrograms per ml. However, it depends on the antigen.You just have to give a try.

About mAb supernatant screening, although some people use undiluted supernatants, I prefer to dilute it at least 1/2 in a soluction containing a good blocking agent (usually 4% milk in PBS) to avoid non-specific background. So I add 50 microliters of the blocking solution and 50 microliters of the supernatant to the well.

About sandwich ELISA, although you can try multiple variants, the starting point could be the following.

1. Coating antibody at 10 micrograms/ml (or 1microgram/well). This is an excess to make sure that you have enough antibody. It can be titrated to know the minimal required concentration, but if you have enough mAb you can start with this without any problem in most of the cases.

2. The antigen is usually used at concentrations below 1 microgram/ml down to a few nanograms/ml. I dont know what are you trying to do with the ELISA. If you want to compare recognition of several antigen variants,  I would try  curves of the antigen (from 1 microgram/ml to  10 ng/ml). This is necessary to see whether the different antigen forms are differentailly recognized in any range.

3. When you say primary antibody, this is not totally clear for me. If you are using an unconjugated  primary antibody and a labelled secondary anti-IgG antibody to detect it, the primary antibody can be used at 1 microgram/ml (again an excess to make sure its enough) or titrated to use lower concentrations. But in that case, the primary antibody and the coating antibody have to be from different species that can be discriminated by the secondary. Is this the situation?

4. On the other hand, if you are using a conjugated antibody to detect captured antigen, working dilution does not only depend on antibody concentration, but also on conjugation efficiency. The only way to determine the working dilution is to titrate the conjugate (1/1000...1/2000........) to find a good signal to noise ratio at some dilution. for this titration ELISA you can coat at 10 micrograms/ml, and use a small curve of native antigen (including blanks without the antigen, of course).

When I say primary antibody, I mean detection antibody.

Sandwich ELISA is to detect native and modified (oxidised) protein through the mAbs that are raised against it (not against whole protein a selected region with modification).

Considering 100ng stock conc for both native and modified the amount of capture & detection antibody required needs to be determined.  

Eg

Modified 100ng 80ng 70ng 50ng 20ng 10ng 0

Native 100ng 80ng 70ng 50ng 20ng 10ng 0

The antigen curve seems to be ok. Coating antibody can be used at 10 microgram/ml and detection antibody (if conjugated to the enzyme) has to be titrated to find a good signal to noise ratio. 

Coating antibody is referred as a capture antibody in the sandwich ELISA. Purified capture antibody 10mg/ml means 10ug/ul. So for 8 wells 100ul/well, 800ul buffer + 8ul antibody from 10mg/ml stock (10ug in 100ul). I was thinking 10ug/ul might be excess cuz during test bleeds and screening we coat the plates with purified antigen (synthetic peptide) at 1ug/well (1ug in 100ul). So why there is an increase in the concentration?

May be I should coat with both 1ug/well and 10ug/well and see the difference.

I said 10 microgram/ml, which is exactly the same as 1 microgram/well. Of course 10 mg/ml (10 milligram/ml) makes no sense, this was not my suggestion. 

If the stock concentration is 1mg/ml then there is 1ug/ul in case of 10ug/ml stock there would be 10ng/ul?  

When protocols mention it as 1ug/well, 1ug in 100ul coating buffer isn't it getting diluted and how do we still call it 1ug/well? Final concentration is changed to 10ng/ul.

And is it wise to use mAbs from two clones A and B raised against same antigen (same modified peptide) in sandwich ELISA? I mean when my capture antibody (clone A) binds to the antigen (modifiedd protein) how does the detetion antibody (clone B) bind again cuz capture antibody has occupied that region or site. In case the detection Ab does bind it implies binding is not specific?  

Note: Detection antibody would be bound by an anti mouse secondary antibody. 

I don't understand the problem to calculate the coating concentration. 10 microgram/ml is the final concentration of the coating solution  that corresponds to 1 microgram/well. Both are exactly the same. One microgram in 100 microliters in the well is the same as 10 micrograms in 1 ml (enough for ten wells). What is exactly the problem with that? My advice is to check carefully the units and their abbreviations until you understand that there is no difference.

If both antibodies have been generated against a very small segment, they cannot be used in a sandwich ELISA (both cannot bind simultaneously) unless this segment is repetitive (present several times per each antigen molecule).

If both capture and detection antibody come from the same species (ie mouse) of course an anti-mouse secondary antibody cannot be used at all. It would recognize the coating antibody independently of the presence of the antigen. That is why you have only two options

1. Use an antibody from a different species as your capture antibody.

2. Conjugate the detection antibody to enzyme or biotin (in the second case you would reveal with conjugated streptavidin.

I think you need to check the ELISA design in general, not only the concentrations.

In the previous comment, it was mentioned 10 micrograms/ml either stock or final concentration was not clear so I considered it as stock concentration because we prepare antigen stock 1mg/ml then dilution (1ul in 100ul) which gives a final concentration of 10ng/ul or 10ug/ml in each well.

I have another option, If we coat with mouse Ab as capture and have my sample (whole protein) then use antibody conjugated with substrate against the protein (anti protein antibody) then it would still be sandwich ELISA? or direct ELISA maybe and are they any flaws in this approach.

I appreciate your response

Yes, you can use a different anti-protein antibody conjugated to an enzyme (not to the substrate) in the sandwich ELISA.