We're performing a long range PCR, longer than 3kb and so far it has failed to work. We've tried DMSO, Mg2+, touchdown approach, different DNA, new primers, different polymerases.
run the pcr in final concentration of 1 molar betaine ( sigma very cheap). It helps the template to melt so improves pcr particularly for high GC templates. Works with all polymerases
The LongAmp Taq (NEB) worked very well for PCR products up to 15 kb in my hands. Have you tried this enzyme already? However, most polymerases like Q5 or Phusion are able to produce products > 5 kb.
Most important is the quality of your template. Have you check your DNA e.g. via NanoDrop? The ratio of 260nm/230nm should be high (at least > 1). An increased amount of dNTPs could be beneficial for really long products, but I do not have any strong evidence for it.
You can also try AccuPrime (Thermo Fischer). It can amplify up to 5 kb. The high fidely one can amplify up to 20 kb.
I'm also trying PrimeSTAR® GXL (Takara) for >20 kb.
I think we can do more suggestions if you show us the protocol that you have used for each enzyme. But like Boas said, the quality of your template is of utmost importance. You need also to check the amount of enzyme and Mg ion in your reaction. And evaluate your cycling conditions.
run the pcr in final concentration of 1 molar betaine ( sigma very cheap). It helps the template to melt so improves pcr particularly for high GC templates. Works with all polymerases